Home

Hyperspectral User Manual

1. Additional Z Profile Cursor Location alue Pixel Locator Point Collection Build Mask Measurement Tool Line of Sight Calculator Arbitrary Profile Transect Spatial Pixel Editor Spectral Pixel Editor Animation 2D Scatter Plots 3D Surfaceview Figure 19 The Profile List is under Tools in the Image tool bar 2 Select a profile X Y or Z In this example of gold nanoparticles the X profile was selected and then the Z profile was selected following the same procedure a When a Profile is selected a cross hair appears on the Image and the intensity plot is displayed b The plot will contain the image intensities along the horizontal crosshair X or the vertical crosshair Y or the Spectral Profile that is contained in the HSI image pixels at the crosshair intersection see Figure 20 TAR cBand YRT GBand 197 B Hand a7 AUD OX HTE Naranai Prania Adea e ouaa amaton esin ile bO 700 0i Wivelength Figure 20 Plot Profiles 3 To view the Plot Key right click in the Spectral Profile Window a Then click on Plot Key 16 4 To display spectra from multiple regions of the image right click in the Spectral Profile window a Select Collect Spectra A new curve is added to the plot each time a new area of the image is clicked see Figure 21 b You can go back to the Zoom Window and select pixels to collect spectra from c Inthe Zoom window clicking increase
2. Memory i A Dutput Resultto File Memory Output Rule Images Yes n Output Result to File Memory Output Rule Images Yes 4 oE C adioa i UF iii Change View 20 2751373 634 OK Queue Cancel Help Figure 14 Window to Preview OK Queue Cancel Hela Preview Figure 15 Spectral Angle Mapper Parameters Window with the Preview window 20 To change the preview area click the change view button located under the classification Preview window The Select Spatial Subset dialogue box opens 21 Click the Image Button the Subset by Image box opens 22 Move the red square to the desired area see Figure 16 23 Then Click OK in the Subset by Image Box and the Select Spatial Subset box 24 Once you have found the desired classification click OK oe Gold nanorod in Tissue 40 air img Di Classification Preview ms 696 x 693 Integer Set Maximum Angle radians Samples i T ol NS C None Single Value C Multiple Values Lines 379 To NL Maximum Angle radians 0 1 0 Full Size 3 964 656 bytes eiben bee 131 072 bytes Output Resultta Fie Memory Subset Using image Map Fie nieve Seco Subset by Image Display H v Output Rule Images Yes uj B Subset by Image E a ey AE i ARE ii Change View 20 275 379 634 OK Queue Cancel Help lt Piot Delete Apply Cancel Samples gt Lines OK Cancel Figure 16 Spectral Angle Mapper Parameters box Select Sp
3. 400 nm 800 nm without the filter 500 nm to 1000 nm with the filter Since the long pass filter does not transmit all of the light past 500 nm it is necessary to also record the light source spectrum when using the filter to do spectral normalization 60 Part 5 Conversions to Reflectance and Absorbance The light absorbance properties of materials can be determined from the absorption spectrum that is recorded as light passes through the material Using HSI the absorption of materials on glass slides can be found and classification of the material based on the absorption property can be carried out Transmitted Light To obtain the absorption spectrum of a specimen on transparent surface use a bright field illumination which sends light directly through the glass slide into the microscope objective wi ied Locate the area of interest on the specimen and center it in the scan area Draw an ROI around the region Save the mean spectrum in a plot After recording the specimen replace it with a blank slide or a region of the same slide that does not contain the specimen Open the ROI tool and use the same region to obtain the mean spectrum of the region Save this to a plot Calculate the absorption spectrum using the spectral math expression a alog float S1 S2 Assign S1 to the spectrum of the blank and S2 to the spectrum of the specimen Choose a new window to create the absorption spectrum The light absorption is maxim
4. IMPORTANT PLEASE READ CAREFULLY Cyt Viva Limited Warranty CytoViva warrants for a period of one 1 year from the date of purchase from CytoViva Inc or an authorized agent of CytoViva Inc the Warranty Period that the unmodified CytoViva adapter and or Dual Mode Fluorescence Module the Equipment when new and subject to normal use and service shall be free of defects in materials and workmanship and shall perform in accordance with the manufacturer s specifications If any component of the Equipment does not function properly during the Warranty Period due to defects in material or workmanship CytoViva will at its option either repair or replace the component without charge subject to the conditions and limitations stated herein Such repair service will include all labor as well as any necessary adjustments and or replacement parts If replacement components are used in making repairs these components may be remanufactured or may contain remanufactured materials Repair or replacement without charge is CytoViva s only obligation under this warranty This warranty is NOT transferable from the original purchaser of the Equipment Limitations Other components of the product package specifically including but are not limited to the light source s light source power transformer s and cord s liquid light guide s optical filters spectrophotometer camera s software microscope part s and motorized stage ar
5. This feature allows you to link both the original image and the classified image to view the same areas 1 Right click in the Original Image Window a Select Link Displays 2 Link Display dialogue box opens a No changes are needed so click OK 3 Now when you move the Zoom box in any of the three images of the original or classified image they move in sync 4 If you left click and hold in any window it will display whats in the opposite window see Figure 23 Dr R Band 187 G Band 117 B Band 47 Gold nanorod in Tissue 4 OX 5 2 Sam Gold nanorod in Tissue 40Xair img Memory3 83 1 Scroll 0 36782 OX SP zomi l x 23 2 Seroll 0 36782 OX 2Z00m 4q OX Figure 23 Linking the two displays 5 To unlink the original image and classification image right click on the original image Then select Unlink Displays 44 Part 8 Overlaying the Classification I mage This feature provides you with the capability to overlay the classification image onto the original image It will enable you to illustrate where the classifcation pixels are in the context of the Original image 1 In the original image window select Overlay from the menu bar a Then select Classification 2 The Interactive Class Tool Input File dialogue box opens see Figure 24 a Select the classification file created from the SAM classifier b For this example select Memory 3 or samples will have a filename c You can look at the Availa
6. Output Files Spatial Resolution C High Low W Subtract Dark Curent Mum Frames 10 Spectral Resolution C High Low Filename for Output Hyperspectral Image Choose Orientation IY Flips Flip Collect Hyperspectral Image i Show Live Preview Figure 2 The control and image scanning control window Part 2 Operating CytoViva ENVI Software The Camera Settings dialog allows the HSI system to recognize your microscope configuration The following controls appear in the dialog box Microscope Objective Magnification This is set to the magnification of the microscope objective that will be used to collect the image example 10X The setting should always match the objective otherwise images will look stretched or compressed from top to bottom Camera Exposure Time Enter a camera exposure time This value is the time that the camera will be exposed to the spectral image on each line of the HSI scan For bright fields of view exposures between 0 1 and 0 5 are appropriate Longer exposures will be needed for low light such as obtained from fluorescent samples 0 5 co The time needed to complete the HSI image collection is proportional to the exposure time that you enter Either use the scroll bar to change the value in 1 second increments or enter the value directly in the text box followed by Return Gain The gain refers to the sensitivity setting on the camera Usually the gain is set low High gain can be used for
7. Plots Set Stack Offset v Auto Scale Axis v O Heia m Reset Range 5 BAA A lt gt New Region Stats Grow Pixel Delete Mickie Clear Plot Filename C Documents lt D a s Optio gt New Window Blanki d a RG Le Weuske fsb hearts Plot_Function gt New Window with Plots 50 D DJD AUD GUL SUD 1000 Select All Hide ROls Basic Stats Min Me Help Wavelength 59 a Annotate Plot 42 ROI tool window Figure 43 ROI Statistics Results Figure 44 ENVI Plot Window 51 5 Click Edit from the menu bar and select Data Parameters 6 The Data Parameters Window opens From here you can see Figure 45 a Label the graph b Rename the selected region c Change the color of the plot see Figure 46 i To change the color right click on the color square ii Move mouse over Items iii Select a color 7 Repeat these steps to for other ROIs that you want to normalize for analysis i Data Paramete ameters T Hiki Available Data Plots Avadoble Data Plots Red Solution n Blue Shio Name Mean Region 1 Red 37 points Name Blue Solution Color Line Style Soid x Color re Shd oF y Items 1 2 06 Elac Thick eems21 40 gt White Items 41 50 Red Thick 1 Nsum 0 SymSize 12 z Symbol None Symbol Line gpl Symbol None Pe Geen Yellow Apply Cy n Le 2 A hillnut a Figure 45 Data Parameters
8. Unfold All Files IR B a gpa g Band 449 986 1235 4uNp 40s air Figure 15 Imaging Open Options 13 Part 8 I mage Enhancements The intensity scale of the image can be stretched or compressed by using several built in options in the Enhance drop down menu on the Image window Stretch Option 1 First select the window containing the HSI image Select the Enhance menu A drop down list shows the built in options see Figure 16 Note stretch options can be selected for any display windows image zoom or scroll 2 The stretch option will be applied to the selected window and also to the other windows However only the selected window will fully adhere to the rules for the stretch option Linear Causes the full range of the display to span between the lowest and highest intensity values of the image This type of image shows the full intensity range without clipping Linear 0 255 Only displays values between 0 and 255 This display is used more for 8 bit images but it can also be used for faint 16 bit images to see fainter objects Linear 2 default displays between 0 and the 2 point on the input histogram This display emphasizes very faint objects Gaussian Equalization and Square Root The stretch options Gaussian Equalization and Square Root perform a nonlinear transformation of the input histogram and the output histogram Square Root is most useful for visualizing both faint and bright ob
9. a gi my pm a Ta Se m E a a i RJE BOG YOO BOG cee 1 Wavelength Figure 6 Spectral curves from three nanoparticles 25 Part 6 Collecting Spectra from Different locations in the Image Multiple spectra can be plotted in the same window This provides an easy way to visually compare the spectral curves of different features Although points for spectra can be chosen from any of the display window it is easiest to use the Zoom window where individual pixels can be identified 1 In the Plot Window right click and select Collect Spectra We have already collected the spectrum of the red particle When Collect Spectra is selected ENVI will collect spectra from any point clicked in all windows I mage Scroll or Zoom Window 2 Now collect spectra of other particles a We will collect the spectrum of the particle that appears yellow b Put the crosshair cursor on the desired particle by left clicking in the zoom window The image moves to place the desired feature at the center of the crosshair c The new spectrum from this location and its label appear together with the red spectrum in the plot 3 Repeat again for a green particle 4 You may end up with three very distinct curves see Figure 7 Note that the color of the spectral curve is assigned by ENVI and does not reflect the color of the feature in the display 5 After collecting spectra a Right click in the plot window b Click plot key This will di
10. aluminum bulb that will proivde light output to about 950 nm Please call CytoViva Inc to order the bulb at 888 737 3130 64
11. b The main ENVI toolbar will open and all of the operations for files image display and data analysis are contained in the drop down menus of the tool bar 2 Inthe menu bar find the CytoViva tab located near the right of the tool bar and begin an imaging session by selecting the menu and clicking on the contents Be sure that the stage controller has been turned on and that it is connected to the computer before proceeding further If the stage is not turned on at this time an error message will be generated and stage movement may occur after the controller is turned on 3 Enter the COM port number see Figure 1 which has been assigned to the automated stage by your computer usually 3 a Click OK EJ Initialize CytoViva Hyperspectral Camera Eg Communication Port for the Hyperspectral Camera COM3 OF Cancel Figure 1 COM Port Window 4 After a few seconds the CytoViva Hyperspectral Camera Settings window appears see Figure 2 The dialogs in this box are used to set up the size and other attributes of the HSI image that will be acquired SY CytoViva Hyperspectral Camera Settings Microscope Field of y en Objective Magnification 100 W Full Feild of View 901 Image Lines Camera Default Bands to Load Exposure Time sec 0 2500 gt Red wavelength 640 460 Gain Low High Green wavelength 550 463 Change COM Port lt Port COM4 gt _Change COM Port Blue wavelength 460 466 Image
12. bands The HSI band and wavelength are shown below the plot when the user clicks on the line The default bands are true color wavelengths red 640nm green 550 nm and blue 460 nm Images will appear with normal colors using the default setting Figure 23 Spectral Profile wavelength display 17 Part 10 Working with Regions of I nterest ROI The Region of Interest ROI tool is used to group pixels of the HSI image for further processing This powerful feature allows preset shapes rectangle ellipse and user defined areas polygons lines and pixels to define a special region of the image ROI s allow groups of pixels to be selected for analysis and plotting ROI s appear on the display as color overlays 1 Right click in the image and select the ROI Tool 2 Expand the ROI Tools window to the right to see all ROI attributes above the box 3 The Tool Selections in this window are used to change the type of ROI add and delete ROIs from the active list and selecting ROIs for plotting ROI Type top tool bar Selects different drawing tools Window Selects the display in which the ROI is to be drawn Select the off button when you are only moving the ROI and not wanting to create an ROI New Region Adds a new ROI to the current list You are able to change the color by right clicking the current color and picking the new color Delete Removes the ROI item in the list Select All Causes all ROI s to be proce
13. keep expressions You can use Restore to bring the expression back for another session 4 Assign the lamp spectrum to the variable S1 If the lamp spectrum is already open in a plot window or saved as a spectral library you will see the spectrum in the list Click on the name to select it Choose to save the results in a new plot Click OK 6 The normalized lamp spectrum is plotted see Figure 58 yn Ma m oe a a m p a a Am mm m ma a RGC Boo Fog Soc Soo 10C Figure 58 Normalized lamp spectrum 58 Part 3 Spectral Normalization of the Specimen Now you have a lamp spectrum for your HSI system Next we will use it to correct your specimen spectra for instrumental effects The correction is to divide the specimen spectrum by the normalized lamp spectrum There are different ways to do this You can correct a single spectrum from a Z profile or correct a mean spectrum from an ROI Alternately you can correct all of the spectra in a data cube 1 Enter or Restore the correction expression float S1 float S2 Add it to the list of expressions and save it to use later Select the expression and click OK 2 Select S1 and make one of the following assignments a To correct a single specimen spectrum select the curve you want from the list It will appear there if it is also in a plot window b To correct a data cube use the file assignment option to browse for the specimen s HSI data file If you do this the entire image will b
14. low light imaging however image quality is better if the gain is kept low and the camera exposure is lengthened If photobleaching is a problem use of high gain will allow a shorter camera exposure and thus a faster scan that limits the amount of bleaching COM Port You will not need to change the COM port so do not use this feature I mage Spatial Resolution Either high or low spatial resolution can be selected for the HSI image data Low resolution is recommended since this mode results in smaller file sizes and increased image brightness Usually brightness is preferred over resolution to get the most information from an HSI image High resolution effectively doubles the spatial resolution however the HSI file sizes are also significantly increased The default resolution is low Spectral Resolution we recommend setting it to low The spectral resolution when high is selected is 1 25nm and when low is selected it is 2 50nm Orientation Make sure only the FlipX box is checked so that the image will be displayed as a direct view instread of a mirror image Field of View The number of scan lines that are needed to achieve a square image scan is calculated by ENVI You may want to record from a smaller area of the field of view To save time and reduce the size of the image file the number of lines can be reduced Change the number of lines by un checking the box enter a new number and press return Remember that more lines will
15. to Subset oa i Band 313 801 1 700 Gold nanorod in Tissue 40airimg 2 just to the right this band corresponds to 800 nm see Figure 11 Then click Add Range Then click OK Band 314 802 4554 Gold nanorod in Tissue 4 0h air img Band 315 803 7415 Gold nanorod in Tissue 4 0h air img Band 316 805 02 70 Gold nanorod in Tissue 4 0h air img Band 317 806 3125 Gold nanorod in Tissue 4 0 air img Band 318 807 5985 Gold nanorod in Tissue 40 airing Band 319 808 8840 Gold nanorod in Tissue 40 air img Band 320 870 7696 Gold nanorod in Tissue 40 air img Band 321 871 4554 Gold nanorod in Titsue 4 0 air img Band 322 812 7410 Gold nanorod in Titsue 4 0 arr img Band 323 814 0265 Gold nanorod in Tissue 40 air ing Band 324 815 31 25 Gold nanorod in Tissue 40 arr img Band 325 876 5985 Gold nanorod in Tissue 40h air ing Humber of tems selected 31 a i Select All Clear Import ASC Figure 11 Select the bands by using the Add Range Boxes 9 The Spectral Subset now shows 312 out of 468 bands are selected see Figure 12 a Click OK again BJ File Spectral Subset Select Bands to Subset Band 295 778 0275 Gold nanorod in Tissue 40 air img Band 296 779 3135 Gold nanorod in Tissue 40Xair img Band 297 780 5995 Gold nanorod in Tissue 40 air i img Band 298 781 8845 Gold nanorod in Tissue 40air img Band 299 783 1705 Gold nanorod in Tissue 40air i img Band 300 784 4565 Gold nanorod in Tiss
16. vertical axis set to 1 a Use Edit menu to change colors and labels see Figure 50 J Spectral Math Result File Edit Options Plot_Function Help Normalized Red and Blue spectra z Wavelength nm Figure 50 Normalized spectra 53 Part 12 Resizing an image Resizing an image allows the user to focus on an area of interest and cut extraneous parts of the image 1 In the ENVI toolbar a Select Basic Tools Click Resize Data Spatial Spectral see Figure 51 The Resize Data Input File box Opens see Figure 52 Select file to resize Click Spatial Subset radio button The Select Spatial Subset window appears Click the Image radio button b C d e f g 7 Resize Data Input File DELER Fie Pemma Classification Transform Resize Data Spatiali Spectral i Select Spatial Subset File S5mq_PCL_60 _1 Dims 696 x 501 Integer Samples To 696 NS 696 Lines f To 501 NL 501 Full Size 697 392 bytes Subset Size 697 392 bytes Sma PCL 60x 1 Figure 51 Opening the Resize window Subset Using Image Map Fie ROI EVF Subset by Image Display 1 Spatial Subset Full Scene Spectral Subset 466 466 Ba OK Cancel Figure 52 Select Spatial Subset window 2 A Subset by Image window appears The whole scan is shown in this window There is also a red box that appears see Figure 53 3 Click and drag the corners of the red box to the preferred size a Click
17. you for purchasing CytoViva products Table of Contents Chapter 1 CytoViva Hyperspectral I maging System I nstallation Part 1 Parts of the Hyperspectral Imaging System Chapter 2 CytoViva Hyperspectral I maging Quick Start Guide Part 1 Starting CytoViva ENVI Software Part 2 Operating CytoViva ENVI Software Part 3 Starting a Scan Part 4 Displaying HSI Images Part 5 Open an Existing File Part 6 Selecting Display Bands Part 7 Opening New Displays Part 8 Image Enhancements Part 9 Plotting Part 10 Working with Regions of Interest ROI Part 11 Plotting and labeling spectra from ROI s Part 12 Saving Images Chapter 3 Extraction of Spectral Profiles Part 1 Opening Test Image Part 2 Acquiring and Labeling a Spectral Profile Part 3 Reading the Data Part 4 Changing the Plot Key Parameters Part 5 Collecting Spectra in the Same Field of View Part 6 Collecting Spectra from Different locations in the Image Part 7 Deleting Unwanted Spectra from the Plot Chapter 4 Saving and Viewing Profiles in the Spectral Library Part 1 Creating an Output Folder Part 2 Saving and Viewing Profiles in Spectral Library Part 3 Viewing the Spectral Library File Chapter 5 Spectral Angle Mapper Classifications SAM Part 1 Automated Comparison of Unknown Spectra with Spectral Libraries Part 2 Using the Spectral Angle Mapper SAM Step 1 Acquiring the Endmember Spectrum Step 2 Saving the Endmember to the Spectral Library Method 1 Using the Mai
18. 0385 Band 9 410 3240 Band 10 411 6095 Band 11 412 8955 a x O Gray Seale AGE Color R E and 188 640 4604 A0100 40 ni r Band 176 550 4625 40 700 400 i p Band 45 460 4655 AU 100 4 0c Dims GE B93 Unsigned Irt BIL Load RGE Display 1 Figure 10 Available Bands List EL Wel sH Wel E Wel Wel E el Oo 11 Part 6 Selecting Display Bands Images opened in ENVI can have different display properties The images can be display in either Gray Scale or Color format Color is the default format 1 ET Available Bands List Slee When the image is first loaded the ENVI automatically picks which image bands to use for red green and blue in the color display The default bands give a true color rendering of the image To be sure that true color is used right click next to the file name at the top of the available bands list and select Load True Color to Current The newly displayed HSI image will revert to the default bands if they are not already being used see Figure 11 Other bands can be selected for red green and blue One reason for selecting new bands is if the HSI image is known to contain no intensity at one or more of the default bands Changing to a different combination of bands will result in false color rendering however it may be desirable to do so in order to create image contrast based on the features of the recorded spectrum Choose new bands by clicking to the left of r
19. 37 687 9596 5mq PCL_60x_1 Band 238 689 2556 5mq_PCL_60x_1 Band 239 690 5522 5maq_PCL_60x_1 Number of items selected f 61 AddRange Select all Clear Import ASCIL 0K cares revo Figure 56 File Spectral Subset window 55 Chapter 6 Quantitative Spectral Analysis In a hyperspectral scan the spectral features are produced by specific light scattering and absorption properties of the specimen materials that are recorded at each image pixel Light scattering is more strongly influenced by surface chemistry where as the light absorption will depend on internal as well as surface properties of the specimen The recorded spectrum also contains features that are properties of the light source that are used to illuminate the specimen and also of optical filters that alter the light reaching the specimen The recorded data actually shows the spectrum of the light from the source going to the sample where it is modified by specific spectral features from the specimen So for example light absorption by a specimen over certain wavelengths may reduce the recorded light intensity at those wavelengths but leave what is essentially a spectrum of the light source at the other wavelengths This occurs when the sample contains light absorbers such as dyes and pigments like melanin that are intrinsic to the sample or if there are dyes that are added to create contrast from sample structures For example the hematoxylin and eosin s
20. CytoViva it should be returned postage paid along with the original dated receipt to CytoViva Inc 300 North Dean Road Suite 5 PMB 157 Auburn AL 36830 Your repaired item or replacement product will be returned to you postage paid In the event the purchaser returns Equipment to CytoViva and it is determined by CytoViva that the Equipment has been returned without cause the purchaser will be notified and the Equipment returned at the purchaser s expense DI SCLAI MER OF WARRANTIES LIMITATION OF LIABILITY THE WARRANTIES CONTAINED HEREIN ARE IN LIEU OF AND CYTOVIVA EXPRESSLY DISCLAIMS AND CUSTOMER WAIVES ALL OTHER REPRESENTATIONS AND WARRANTIES EXPRESS OR IMPLIED STATUTORY ARISING IN THE COURSE OF DEALING OR PERFORMANCE CUSTOM USAGE IN TRADE OR OTHERWISE INCLUDING WITHOUT LIMITATION IMPLIED WARRANTIES OF MERCHANTABILITY TITLE INFRINGEMENT OR FITNESS FOR A PARTICULAR PURPOSE WHERE NOT PROHIBITED BY LAW CYTOVIVA WILL NOT BE RESPONSIBLE FOR ANY CONSEQUENTIAL OR INCIDENTAL DAMAGES RESULTING FROM THE PURCHASE USE OR IMPROPER FUNCTIONING OF THIS EQUIPMENT REGARDLESS OF THE CAUSE SUCH DAMAGES FOR WHICH CYTOVIVA WILL NOT BE RESPONSIBLE INCLUDE BUT ARE NOT LIMITED TO LOSS OF REVENUE OR PROFIT DOWNTIME COSTS LOSS OF USE OF YOUR EQUIPMENT COST OF ANY SUBSTITUTE EQUIPMENT FACILITIES OR SERVICES OR CLAIMS OF THIRD PARTIES FOR SUCH DAMAGES CYTOVIVAS LIABILITY WILL NEVER EXCEED THE PURCHASE PRICE OF ANY DEFECTIVE PRODUCT OR PART Thank
21. ES Part Number 1055132 Nexterion Glass B Ultrasonically clnd COVER SLIPS Part Number 1098576 Nexterion Glass D type 1 D263 Custom Coverslips 62 Glossary Image File F This is the file of the original image scanned with the spectrophotometer Spectral Library File This file contains the different spectra that the user will collect from the scanned image Rule I mage File This file contains the Rule image of the sample Classification SAM File 1 This contains the classification SAM Image of the spectra that matched the endmemebers Rule I mage The rule image is the displayed grayscale image showing the degree of matching between the known and unknown spectra from each image pixel The Rule image is shown on a scale of light to dark the relative degree to which unknown spectra in each image pixel match the known spectra Spectral Angle Mapper Classification SAM an automated procedure for determining if a known material is present in the input image and the pixels containing the material SAM accomplishes these tasks by comparing unknown spectra in Hyperspectral imagery with known laboratory spectra for the material in question Classification SAM I mage this image shows the location of pixels whose spectra match that of the endmember Header File A separate file created with the Hyperspectral image file with the same name This file must be copied with the image file when data is transferred The heade
22. NVI IDL Software by double clicking on the ENVI_IDL icon on the desktop 2 Click on File a Select open image file b Select CytoViva Test Images These test images are preloaded onto your ENVI software c Select AU nanorods 40x img 3 Load the image into a new color display see Figure 1 4 Set the display enhancement d Using the Image Menu click on the Enhance drop down list e Click on Linear 0 255 3 Observe that there are small particles spread over the image and that the particles are mostly red yellow and green The colors viewed in the display are determined by which bands are chosen to represent red green and blue in the available bands list Although the particles appear to have different color properties we need to see if the spectra of the different colored particles are similar or very different 4 Inthe figure above the Zoom window area is marked by a red rectangle by 1 R Band 187 G Band 117 B Band 47 Purd E BF Fie Overlay Enhance Tools Window Figure 11 AU nanorods 21 Part 2 Acquiring and Labeling a Spectral Profile Spectra can be plotted from image features that are selected by the cursor This action uses the Z Profile feature 1 Enlarge the Zoom window so that it contains many objects of different color see Figure 2 2 Increase the zoom factor so that individual pixels are observed by using the and tabs in the zoom window 3 To turn on and off the crosshairs use the button to the
23. OK to exit Subset window b Click OK to exit Spatial Subset window c Click OK to exit Resize Data window d Save file accordingly and the newly resized image will appear Samples 298 Lines 212 OK Cancel Figure 53 Subset by Image window 54 Part 13 Spectral sub setting an image Spectrally sub setting an image allows the user to focus on specific wavelengths of light This is useful in removing second order harmonics and other known reflective contaminants 1 Inthe main ENVI toolbar a Select Basic Tools b Click Resize Data Spatial Spectral see Figure 54 c Choose image to resize in the Select Input File list d Click the Spectral Subset radio button see Figure 55 EJ ENVY y Resize Data Input File Select Input File Image output emaple Amg PCL 60x 1 File Pemma Classification Transform Resize Data Spatiali Spectral Figure 54 Opening the Resize window Spatial Subset Full Scene _ Spectral Subset 465 466 Bands Figure 55 Resize Data Input File window 2 The File Spectral Subset window appears 3 Select the range of spectra to become the subset by clicking and scrolling see Figure 56 To select non contiguous values hit the Ctrl key 4 Click OK to exit Spectral Subset window a Click OK to exit Resize Data window b Save file accordingly and the newly sub setted image will appear File Spectral Subset Select Bands to Subset Band 236 686 6642 5ma_PCL_60x_1 Band 2
24. Part 1 Creating an Output Folder When doing spectral analysis on scans you will need to create several different types of output files which could include but are not limited to e Spectral library files e Header Files e Spectral Angle Mapper SAM Files These are files of your spectral data that you will want to access later to compare against new samples The Output folder will contain files from scanned images 1 To create your output folder choose a location within your images folder 2 Click on Create New Folder 3 Label the folder Output see Figure 1 Image Folder File Edit Yiew Favorites Tools Help CES gt P Search n Folders EHK Address ie C Documents and Settings Chuck Ludwig Desktop Image Folder File and Folder Tasks Output D Make a new Folder a Publish this folder to the Web E2 Share this Folder Other Places amp Desktop E My Documents O Shared Documents 4 My Computer J My Network Places Details Figure 1 Output folder for saving image work in the Images Folder 28 Part 2 Saving and Viewing Profiles in the Spectral Library 1 After you have created your output file refer to Part 1 Creating an Output Folder the red yellow green and blue particle spectra collected earlier can be saved to a spectral library From the spectral profile window select File a Then select Save Plot As b Then select Spectral Library You will see the Output Plots to Spect
25. Window Figure 46 Changing graph color Normalizing 1 Using the ENVI Plot Window that has the spectra to be normalized open see Figure 47 r gt VENYI Ploy Window Blue Solution Ys Rec a00 ra Wavelength nm Figure 47 Spectra to be normalized 52 4 2 OF Cancel Help In the main ENVI toolbar a Select Spectral Then select Spectral Math The spectral math window will appear see Figure 48 Enter the normalizing expression float s1 max s1 Click OK A Variables to Spectra Pairings window will appear In the Available Spectra list click the curve to be normalized Click the double arrows to change Output Result to New Window Click OK see Figure 49 j Repeat for all desired spectra sa m7oan0c Sy Spectral Math Ed El Variables to Spectra Pairings Eg Previous Spectral Math Expressions Exp float s max s1 floats masis Variables used in expression 51 Aed Solution Ay ailable Spectra list Min Region 1 Red 37 paints Stdev Region 1 Red 3 points Save Restore Clear Delete Mean Region 1 Red 37 points Enter an espressioni Max 1 1 Red 37 float st frnax s1 Red Solution Blue Solution Add to List Map Variable to Input File Output Result to Hew Window ttl Figure 48 Normalizing expression in Spectral Math window OF Queue Cancel Help Chess Figure 49 Variables to Spectra Pairings window A new graph will appear with a
26. atial Subset box Subset by Image 38 25 The SAM files are now added to the Available bands list Now there are two images added to the top of the available bands list 26 To load each image in a new display right click on the Rule Image Then select Load Band to New Display 27 An image will open 28 Then right click on the SAM image 29 Then select Load Band to New Display 30 Another image will open 31 You should now have the original image the Rule Image and the SAM image shown on your screen For future SAM classifications you can save Output Results and Endmember Spectra by choosing the File option rather than the Memory If you close out the ENVI software this file will be lost since it was saved to memory 39 Part 3 Re Loading mages from Available Band Lists Window In some cases you may load one image over another by accident This usually happens by clicking Load Band to Current Display rather than Load Band to New Display If this happens it will load the new image over the original image You will need to reload the old image to a new display 1 To reload the original image see Figure 17 scroll through the Available Bands List to find the original image In this exercise the image name is Gold Nanorods in Tissue 2 Click on the filename to highlight it Click Display and select New Display 4 Click Load Band to again display the original image a Available Bands List KIBI File Op
27. ble Bands List to help determine which one is the SAM image d Click OK 3 The 1 Interactive Class Tool dialogue box opens a Click on the tab beside the red box marked AU NR Tissue see Figure 25 9 Interactive Class Tool Input File Select Input File File Information File Memory ltem Memory3 Dims 696 x 693 1 B50 On 4 Size Byte 482 328 bytes nanorod in Tissue 40air img File Type ENYI Classification urdue 4U nanorods 40 ima Sensor Type Unknown r Byte Order Host Intel 1 Interactive Class Tool Farah eee File Edit Options Help Wavelength None Upper Left Corner 1 1 Active Class E Unclassified Description Spectral Angle Mapper Classification Result Tue Oct 21 16 32 13 2008 A On i Unclassified V Or a Au NR Tissue Figure 25 activating the overlay feature x OK Cancel Open Figure 24 Selecting image to overlay 4 Now you will see an overlay image of your classified picture and the original image see Figure 26 7 1 R Band 187 G Band 117 B Band 47 Gold nanorod in Tissue 4 O X b 7 1 Scroll 0 36782 O X 5414 Zoom 4x Figure 26 View of the images overlayed 45 Part 9 Merging Colors in a Classification Image Sometimes multiple spectra are used to classify particles in a sample This will give an overlay that has multiple colors It will allow you to combine the multiple colors into a single color 1 Open the image to be classified see Figu
28. cale of the Rule image is often reversed to create this effect Additionally a criterion for the match between known and unknown spectra can be given When this is requested SAM produces a binary classification image showing locations of pixels with spectra that are within the criterion The SAM method works by comparing unknown spectra from each image pixel with the known spectrum for example in a spectral library at each of the N wavelength bands recorded in the image SAM first determines a vector in N dimensions that represents the distance from the origin dark to the light intensity recorded in each band of the unknown spectrum The direction of this vector in N dimensional space is used to define a unit vector representing the unknown spectrum The same procedure is performed for the known spectrum SAM then determines the angle between the two unit vectors The best spectral match occurs when the angle between these vectors is the smallest The selection criterion that was used to make the classification image is actually a threshold for the spectral angle Image pixels with spectral angles that are smaller than the threshold are classified as containing material belonging to the known spectrum In plot shown below see Figure 1 the unknown spectrum was not classified as belonging to the known spectrum since the angle turned out to be larger than the criterion allowed The angle of each pixel is output to the Rule image Because the m
29. can also select Plot Parameters from the Image Bar Menu to change the X amp Y Axis Title of the Plot and Background and Foreground color 1f you don t click apply the new spectral curve label will not be changed Spectral Profile R G Boo Yoo Boo d 100g Wavelength Figure 5 Labeled plot 24 Part 5 Collecting Spectra in the Same Field of View 1 In the Plot Window right click and select Collect Spectra This has just collected the spectra of your red particle in the sample When the Collect Spectra is selected ENVI will collect spectra from any point clicked in all windows I mage Scroll or Zoom Window 2 Now to collect spectra of another particle a In this case we will collect the spectrum of the particle that appears yellow Put the crosshair cursor on the desired particle by left clicking in the zoom window and moving to desired location b The new plot and its label appear together with the red spectrum and label Repeat again for a green particle You may end up with three very distinct curves see Figure 6 Finally change the names and colors of these new spectra so they can be recognized later Refer to Part 4 Changing the Spectral Curve Label name _ SS O ae E SE D O B j i i w t e a u a i i k i UPE E 7 i fp i F lt 4 F EAT Spertral Profils iu orme A fz E a eu epee aa of ie i a a HE raf en al Spectral Profile m ty m
30. ced by light scattering in the medium around the particle The cell is gray only because it scatters more light than the background and hence its spectral curve has a greater amplitude even though it does not match the signature of GNR Figure 18 shows the rule image obtained by SAM L TPA Fiabe CAU NE tianun Memory EN OF Jom 4x Figure 18 SAM rule image from gold nanoparticles in cell 41 Part 5 The Classification SAM I mage The classification image shows the location of pixels whose spectra match that of the endmember You will see that using a value of 0 1 for the maximum angle causes only pixels in the selected particle to be classified as a match for the endmember spectrum The zoom window rectangle surrounds this particle as shown in Figure 19 In the zoom window directly below the main image you can see that three pixels were classified The rest of the pixels within the particle contained sufficiently different spectra as did those of every other particle and they were rejected as matches for the endmember This outcome suggests that the spectral angle chosen for the classifier is too small Figure 19 Classification image using a low maximum angle in SAM 42 Part 6 Re Classifying the SAM For this example we believe other gold nanorod GNR particles contained in the image have spectra that are similar to the spectra collected for the first particle Thus we will try to add more
31. croll This window contains the full HSI image The size and position of the Image window is marked with a red square The red square can be moved to enter different parts of the full HSI image into the Image window Zoom A sub region of the Image window appears in the Zoom window The red square marks the shape and the position of the zoom area in the Image window This window can also be resized and moved within the Image window The zoom scale can be increased or decreased at the lower left corner of the window Here is the ENVI display of an HSI image of gold nanoparticles using the color default see Figure 9 0 1 R Band 187 G Band 117 8 Band 47 aut9 O x Figure 9 Color image set after scan includes Image Scroll and Zoom displays 10 Part 5 Opening an Existing File 1 Using the Main Menu bar previously recorded HSI images can be opened from the File Open dialog The HSI image is automatically loaded into three displays Image Scroll and Zoom windows 2 Upon opening an image file the file name appears in the Available Bands List see Figure 10 3 Under the name the individual HSI image bands are listed showing the associated wavelengths Individual bands can be selected and displayed Available Hands bist El AU100 40s0il 2 2 250ms A O Band 1 400 0385 Band 2 401 3245 Band 3 402 6100 Band 4 403 8955 Band 5 405 1915 Band 6 406 4670 Band 7 407 7525 Band 8 409
32. ct file then click okay J Interactive Class Tool Input File Select Input File File Information MM_60x_2 MM_60x_2_slf_on_Ngel MM_60x_2 Nael MM_60x_2 MM_60x_1 clump _slf_on_Ngel MM_60x_1 Nael MM_60x_1 MM_60 _1 OK Cancel Open Previous File gt New File Spectral Library ROI File EVF File Figure 29 Interactive Class Tool Input File window 1 f the classification file is not in the Select I nput File list then click Open and select New File 46 8 An Interactive Class Tool window will appear 9 To show the overlay a Right click on the box next to any colored square see Figure 30 b In the Interactive Class Tool window click Options c Select Merge classes see Figure 31 gt TAT Bifand 200 Gand TZE B Band Ta ieai MM 6Oe 1 t7 1 Interactive Class Tool b7 1 Interactive Class Tool File Edit enisiss Help File Edit Options Help Class distribution Active Cla Active Clase fl Unclassified Edit class colors names Merge classes A t BB Unclassified D Rasys Mean For active class Mean For all classes MB fx s04 7 272 E kasy lv On Si alee Hot abate rack WE ixsc0v 257 E povo Fe Stats For all classes B kas E eav ah ave W On Calculate Covariance with Stats O KaYa E pav ies Y On associated stats data file EN Class Transparency W On Hide Window Figure 30 Overlay Image with Interactive Class Figure 31 Options drop down menu T
33. ction dialog with entry from label dragged from plot 13 You will see the Spectral Angle Mapper Parameters dialog box open 14 For Set Maximum Angle radians click on single value 15 Keep the Maximum Angle radians at 0 1 We will use this threshold first for the spectral angle 16 For Output Result select memory This will send the classification image to the available bands list The Output Result image can also be saved as a file The filename for this image should be chosen as filename_SAM dat 17 For Output Rule Image select Yes 18 For Output Result select memory This image can also be saved in a file using the filename filename_SAMrule dat 37 19 If you want to Preview the Classified Mapped Image click Preview see Figure 14 from the Spectral Angle Mapper Parameter window This will show you what pixels are going to be classified before saving 20 After you click Preview the preview window is added to the SAM Parameter window see Figure 15 The overview window ONLY shows the pixels of a section of the entire image that are classified by the SAM Unclassified portions of the sample will not be shown 7 Spectral Angle Mapper Parameters EJ E Spectral Angle Mapper Parameters Set Maximum Angle radians Saan aie cadens Classification Preview C None Single Value Multiple Values Hore i gt Spena T Muhole Vaiss Maxi Angle radi 0 10 Sree Maximum Angle radians 0 1 0 Output Resultto File
34. ctra will be displayed Right click in the plot window Select Plot Key Right click again and select Options Select New window Blank Left click on the legend for the white curve that contains the mean spectrum Drag the name to the blank plot It will show the lamp spectrum by itself Save the plot as a spectral library so you can do further work with the lamp spectrum see Figure 57 a on JSCOUL z Tatafa Kj 2 QD J D 1090 ats Ae A oTates on avs S06 6006 00 800 906 10i Wavelength nm Figure 57 The spectrum of the halogen lamp acquired with a Pixelfly camera and spectrograph on CytoViva HSI The spectrum shows the shape of the halogen lamp spectrum Note This contains effects of the camera and spectrograph upon the actual halogen light spectrum emitted from the light source 5 Part 2 Spectral Normalization of the Lamp This procedure removes the effect of the lamp spectrum so that sample features are seen more clearly 1 First set the value of the highest point on the lamp spectrum to unity If you recorded a new lamp spectrum go to Basic Tools located on the main ENVI menu select Spectral Math The halogen lamp spectral library provided by CytoViva has already been adjusted this way 2 Type inthe expression float S1 max S1 and add it to the list of expressions Select the expression and hit OK 3 Use the Save button to save this expression to a special folder where you will
35. e Preview by pressing Collect Hyperspectral Image and review the new intensities When the Preview shows good data press Capture to start the scan If dark subtraction was selected the following message appears see Figure 6 Do not press OK yet Information l Preparing to measure dark current Turn off illumination Figure 6 Dark Subtraction Window At this point divert the light path from the HSI camera to the microscope eyepiece by pushing the slide bar in all the way on the trinocular head Then press OK Do not turn off or change the brightness of the illuminator A recording of the dark image will be made from the requested number of dark frames The dark image that will be used is the average of the recorded dark frames 13 Once the dark currents are captured the following message to begin the scan appears next see figure 7 Do not click Yes or No yet Question E Captured 10 dark current images and calculated mean Please open aperture For slide image capture Do vou Wank to view the dark current image Yes Mo Figure 7 Dark Current Message Window 14 At this point divert the light path from the eye pieces to the HSI camera by pulling the 15 Slide bar all the way out on the trinocular head Select No It is not necessary to view the dark image The HSI scan will begin and the acquisition progress bar appears see Figure 8 The stage moves to the beginning of the scanning area and then
36. e collected spectra individually or together 1 R In the Available Bands List Window right click on the file name of the desired spectral library file Choose Spectral Library Viewer In the Spectral Library Viewer select the red spectrum A plot window opens with the spectral curve Select the other spectra in the Viewer They are added to the plot Note that curves saved in the spectral library are also referred to as endmembers in automated spectral classifications The use of endmembers will be discussed next 30 Chapter 5 Spectral Angle Mapper Classification Part 1 Automated Comparison of Unknown Spectra with Spectral Libraries The Spectral Angle Mapper Classification SAM an automated procedure for determining if a Known material is present in the input image and locating which pixels contain the material SAM accomplishes these tasks by comparing unknown spectra in hyperspectral imagery with known spectra for the material in question The degree of match between the unknown and known spectra from each image pixel is displayed in a grayscale image known as the Rule image The Rule image shows on a scale of light to dark the relative degree to which unknown spectra in each image pixel match the known spectrum The scale is reversed from convention so that best matches produce the darkest pixels Since by convention high correlations are scaled toward the bright end of the scale in image processing the grays
37. e corrected 3 Choose a new window to plot the corrected spectrum A new filename for the corrected data cube can be chosen 4 Inthe assignment window select the normalized lamp spectrum for S2 see Figure 59 and 60 Recarded AuNP 5 S000 200G ago ma r I 1500 OOR 1500 TOU 11EG H S k k ee S00 6900 76O 800 900 1000 BOG 600 FOO 800 900 1000 Wavelength nm Wavelength nm Figure 59 The uncorrected spectra of Figure 60 After correction for instrumental two nanoparticles effects Note The correction results in noise where the uncorrected spectrum had no intensity This noise is not related to features of the particles 59 Part 4 Correction for second order diffraction The range of the spectrograph used with CytoViva is from 400 nm to 1000 nm Since the spectrograph is a grating type the spectrum starting at 400 nm also starts again at 800 nm because of the second order diffraction of the grating This means that there are spectral values between 800 nm and 1000 nm that mirror the values from 400 nm to 500 nm Values between 400 nm and 800 nm are always obtained from the fundamental wavelengths i e the first order diffraction pattern and are never contaminated by the second order To obtain true values out to 1000 nm we recommend using a 500 nm long pass filter to block the light between 400 nm and 500 nm which will produce interference past 800 nm This gives two free spectral ranges
38. e cursor at any place on the image corresponding to a spectral wavelength causes the intensities of all the points across the image at that wavelength to be plotted Placement of objects The features on the specimen that will be scanned are first moved into the field of view of the eyepiece and centered at the intersection of the cross hairs using the stage joystick With the specimen positioned thus the stage can be moved forward and backward small amounts to center the desired area on the crosshairs Note any places where the features are especially bright and move these so that they intersect the horizontal crosshair The spectra of these same objects will appear in the Preview screen Note that the live frames are flipped reversed about the X axis The spectral images below are from two bright objects that appeared directly above the horizontal 5 crosshair Focus Typically if the objects are in focus in the eyepiece of the microscope then they are in focus to the HSI camera Often the thickness of the specimen makes it difficult to properly adjust for best focus This can be done more easily with the Preview image If focus is poor the spectra will appear blurred see Figure 3 When focused correctly the spectra become sharp see Figure 4 Focusing with the Preview window insures that the HSI image will be sharp Sl CytoViva HSI Microscopy Preview SEE SI CytoViva HSI Microscopy Preview Capture Cancel Captu
39. e system are necessary as well 1 Use a KimWipe or lens paper to wipe off any excess oil from the CytoViva condenser and the objectives 2 Power off the light source 3 Turn the Dage camera off by unplugging the power cord connected directly to the top of the camera 4 Optional turn the stage controller unit off using the switch on the back of the controller box 5 Cover the microscope with the CytoViva microscope cover or similar 6 Use cleaning solution about every 2 3 months to clean the CytoViva condenser and the objectives Letting the immersion oil sit and dry on the condenser and the objectives can cause a gum like film which can cause damage to the lens Part 2 CytoViva Supplies Below listed are supplies that are recommended when using the CytoViva system for imaging and clean up Fisher Scientific Pre cleaned slides 1 x 3mm plain glass slide Part 12 544 1 Pre cleaned coverslips 17mm thickness No 1 slides two sizes below 18mm x 18mm Part 12 542A 55mm x 24mm Part 12 544 18 Cargille Immersion Oil Type A 16 fl oz Part 12 368B TexWipes good wipes for cleaning Part 18 315A Fisher Brand Optical Lens Cleaner use for cleaning Part 22 143 974 Schott North America I nc Contact Lynn Larcheveque lynn larcheveque us schott com 502 657 4412 Mention that CytoViva referred you Ultrasoncially cleaned slides and coverslips recommend for best imaging results SLID
40. e warranted based on the individual original manufacturer s warranties and policies There is no warranty whatsoever on the contrast filters bulbs or the coil on the motorized stage purchased as part of the Equipment This warranty does not cover circumstances beyond CytoViva s control breakage or a malfunction that has resulted from improper or unreasonable use or maintenance accident tampering misuse neglect improper installation modification improper maintenance or service cleaning procedures shipping or repacking of Equipment or service or parts to correct problems where such service or parts are performed or provided by anyone other than CytoViva or an authorized agent of CytoViva service required as the result of unauthorized modifications or service misuse or abuse failure to follow CytoViva s operating maintenance or failure to use items supplied by CytoViva This warranty is also void if the light source light source power transformer and cord or liquid light guide is not used in accordance with the original manufacturer s instructions recommendations or documentation Warranty service will not be provided without a dated proof of purchase Please return the Warranty Registration Card together with a copy of the original receipt within thirty 30 days of purchase It is the purchaser s responsibility to return the Equipment to the authorized agent from whom it was purchased If the Equipment was purchased directly from
41. ear the left of the tissue section see Figure 2 Use the box in the Scroll Window to locate the area of the cluster of particles Then use the zoom box in the Image window to better locate focus the cluster of particles so that they appear in the zoom window 7 1 R Band 187 G Band 117 B Band 47 Gold D Figure 2 Gold nanorods inside rectangle in cell cluster 6 Using the Zoom Window select a clear distinct particle using the crosshair cursor see Figure 3 We will use this particle to represent the GNR that are distributed throughout this image 7 1 Zoom 5x Figure 3 Crosshair centered over particle in Zoom Window 32 7 Now to plot the spectrum of the GNR a Inthe Zoom Window right click and select Z Profile Spectrum b Before acquiring the profile right click in the plot window and set the Z Profile Average Window to a size that fits within the borders of the particle Here a 3x3 rectangular pixel average is selected 8 Move the crosshair cursor over the center of the particle as in Figure 3 and left click to obtain the endmember spectrum that will be used to classify the image 9 Now give the ploy key label a new name Refer to Chapter 4 Part 4 Changing the Plot Key Parameters that allows you to find it in the future Here we chose AU NR tissue see Figure 4 iolile Gold nanorad in Tes Spectral Profile a Tm Ti mi ym ey mr o E ee BOG BOG YOR Boo j 1606 Wavelength Figu
42. ed green or blue Then scroll to the desired band in the window above and click on it After changing the image band associated with one of the colors ENVI automatically moves to the next color When you are done selecting the bands press the Load RGB button see Figure 12 When Gray Scale is selected only one band is used to display the image EJ Available Bands List Ses File Options File Options HE SU open File in ENVI Zoom 4dd Selected File to Memor Close Selected File o Band 451 363 0254 O Band 452 990 4773 Edit Header O Band 453 991 9258 ate bebe O Band 454 995 3825 m Band 455 994 5367 Band 456 996 2912 Band 457 997 7463 Band 458 999 2020 Band 459 1000 2942 O Load Trie olor Load CIR m O Load Default RGE m Fold All Files C Gray 5 Unfold all Files Gray Seale AGE Color Dime TR Band 195 638 3976 AuNp 40s air g Band 124 548 5525 AuNp 40x air Wri Band BO 458 4062 AuMp 40 air ESE 399 Integer BIL Load RGB No Display Load RGE Display 1 a Figure 11 Load true color menu bar Figure 12 Automatic RGB bands Sop Band 449 986 1235 AuNp 40 ai Band 451 989 0254 AuNp 40s air og Band 449 996 1235 AuNp 40x air Dims ESE 399 Integer BIL 12 Part 7 Opening New Displays 1 Additional display windows can be opened using the No Display button on the Available Bands List see Figure 13 qd Select New Display a bla
43. eft corner of the plot see Figure 4 Read the information from specific locations on the spectrum by left clicking and holding the button down as you move in the plot window A line cursor follows your movement and the band wavelength and intensity are updated in the lower left In this example Figure 4 the data show that the cursor is at band 139 the wavelength is 577 46nm and the intensity of the spectrum is 373 7 units ET 4 Spectral Profile Gold nanorod in Tiss Seles File Edt Options Flot_Function Help Spectral Protile 1 oe ee oo PAE E 19C r i y pa bay il A rir as aie FT id rata Pa aie m m m a m m m m a a m ee ine SL ADU fbU rat a Gi 10 139 577 468 373 68750 Wavelength Figure 4 The Spectral Profile with vertical cursor and data readout 23 Part 4 Changing the Plot Key Parameters 1 In the plot window right click 2 Select Plot Key you will see the spectral curve label name in the right margin 3 To change the Plot Key names line color and style so that can be identified later go to the plot window and on the image menu bar a Select Edit b Then select Data Parameters c Inthe edit window select the label and below enter the new name in the box below d To change the plot color and line style right click on the box and select the desired line color For this example we named the particle spectrum Red e Click Apply and then close the edit window see Figure 5 4 You
44. etected in the image The control of spatial and spectral resolutions work the same way Spectral Resolution The minimum separation in wavelengths that can be detected The low spectral resolution setting is equal to half that of the Hi setting Therefore if the hi resolution were 1 5 nm the low resolution would be 3 nm 63 Trouble Shooting What is the number for Technical Support 1 888 737 3130 After the scan has finished the I mage on the screen is Black One of the slide bar wasn t opened or closed at the proper time Refer to the Manual Chapter 2 Quick Start Guide Part 3 Starting the Scan numbers 6 amp 7 The scanned image is out of focus Before you scan your sample be sure to double check that the sample is in focus Sometimes the slightest movement can make the sample move When I rescan an image the original image is still showing Go to the Available Bands list and choose close all files and then in the Available Bands list and open the new image you just scanned You can refer to the Manual in Chapter 2 Quick Start Guide Step 3 Starting the Scan and number 8 which provides more detail There is a black area at the right or left end of my scanned image This is due to the slit not being pushed in all the way on the spectrophotometer Call Technical Support can t get spectra past 750nm CytoViva offers a different bulb for the Dolan Jenner than the dichroic bulb provided The bulb is an
45. ethod is based only on the direction of these vectors and not on recorded light intensity the SAM classification is insensitive to the illumination of the sample in the recording Therefore it is important that dark values are removed from the data before using SAM Since sample illumination varies over the field of recording and between images the SAM tool is a highly useful classification method Unknown Threshold F Training class pixel Band 2 Band 1 Figure 1 Threshold for classification of pixels based on spectral angle 31 Part 2 Using the Spectral Angle Ma r SAM This example examines an image of gold nano rods GNR in tissue The object of this exercise is to automatically locate the GNR objects within the image First the known spectrum for GNR is acquired and saved in the spectral library Spectra that are used as a signature for identifying a material are called endmember spectra The SAM classification feature then is used to find the locations of objects matching the spectrum of GNR in the image Step 1 Acquiring the endmember spectrum if 2 i In the Main Menu Bar use the Open Image File to load the input image Click on the file Gold Nanorod in Tissue 40Xair img preloaded in your software a To optimize the display contrast click on Enhance in the Image Window b Then click Linear 0 255 this will enhance the display contrast and brightness Move the Zoom area over the cluster of particles n
46. his example for Output Result select Memory b Your spectral library is now present in the Available Bands List as data in memory EJ 1 Spectral Profile Gold nanorod in Tiss fe felix ie Edit Options Plot Function Help Input Data P Raa Save Plot s ASE i Spectral Library IDL Variable Print Cancel PostScript Image File Aan ree L Li 16 POD 706 BOG 16060 Wovelength Figure 7 Saving Endmember to Spectral Library File 4 To view the endmember plot a Right click on Memory in the Available Bands list see Figure 8 b Then select Spectral Library Viewer This now shows the plot for the endmember s in this case we only have one see Figure 9 Available Bands List TBR File Options Memon Spectral Library Viewer o Si a ge Save Selected File to Disk Close Selected File Edit Header Quick Stats Fold All Files Unfold All Files TITS zgo B B B o B o B m B B B B Gray Scale RGB Color Selected Band Spectral Library Memory1 F ROC gob 10C E EAA Figure 9 Plot containing the spectrum of the Endmember Load Band Display 1 Figure 8 Using the Spectral Library Viewer to view Endmember 35 Step 3 Performing the SAM Classification We will use SAM to create a Rule image that shows how well each pixel of the input image matches the endmember spectrum Then we will create binary images sh
47. inimize when you navigate away 48 3 Open the spectral library see Figure 36 a Inthe ENVI task bar select File click Open Image File Select spectral library file to be edited A Spectral Library Viewer window will open Click on each spectra to create a spectral library plot In the Spectral Library Plots window select Options Click on show plot key and all spectra will be listed to the right of the spectral plot pang AJ Spectral Library Plots ex 5 Spectral Library Viewer File Options Library KF140711C_slf 450 bands Wavelength 400 243195 to 1000 711121 Nanc 203 Y 121 202 Y 121 x 202 Y 120 x 202 Y 119 x 202 Y 118 rable Wavelength Figure 36 Spectral Library Viewer window right used to create the Spectral Library Plot left f Right click on the spectra to be removed g Then select Remove X Y for all spectra chosen see Figure 37 mtg 1 b2o a Fn B P 4s P 429 Y1 Remove s204 il21 w Plot Key Stack Plots Set Stack OFFset w Auto Scale Y Axis Reset Range A o Clear Flot Tod Figure 37 Selection of spectra from Plot Key to be removed 49 4 Save the file as a Spectral Library see Figure 38 a Click File Select Save Plot As Click Spectral Library Click Select all Items in output plots window see Figure 39 Then click OK Click Choose in the Output Spectral Library window see Figure 40 Enter file name in Output Filename window click Open C
48. is is done by making all the intensities equal to 1 Mean Spectra Analysis 1 Inthe Image Window menu bar select Tools see Figure 41 a Select Region of Interest b Click ROI tool by 2 R Band 199 G Band 127 B Band 52 red_soln_60x_1 File Overlay Enhance BESSI Window Link gt Profiles d Polarization Signatures d gt Region Of Interest ROI Tool Color Mappin gt PERN Restore Saved ROI File Figure 41 Opening the ROI tool 2 Highlight the Region of Interest needed refer to Chapter 2 Part 10 Working with Regions of Interest ROIs a Then Click the Stats radio button see Figure 42 3 The ROI Statistics Results window appears a Right click on the graph Click Plot Key Right click on the graph again Select Options Click New Window Blank see Figure 43 f Anew ENVI Plot Window opens see Figure 44 4 To load the mean spectra into the New ENVI Plot Window a Click and drag the Mean Region of Interest from the ROI statistics window into the new ENVI Plot Window anoo Note Multiple graphs can be loaded into this window for comparison by dragging and dropping the required data T ROI Statistics Results red_soln_60x_1 File Options 6 ENVI Plot Window E7 2 ROI Tool File ROL TYS Oprig Help Select Piot Clear Plot o Window Image C Scroll Zoom Off Daka Parameters Plot Parameters Mean Region ROI Name Color Pixels Poly Data Yalues v Plot Key 1000 Stack
49. ive Class Unclassified A BB Unclassified Von W ker BB PZA Mon E kya Epoa won B kava Epoa vo B Rava E parva Po B parm B parm vo B para Epa wom W pao Yy Figure 34 View of the Overlay Image b Click on Class Distribution see Figure 35 Classification Distribution SEE File KF140711C_slf_on_BSA_sol_SAmM Class Name Npts Bet 172960 99 802 0 198 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 000 PD DD DD SSS kas kar kar haz an Led at el ee Leh a O a eee ee iras i Grani aai ieas ee ee ee ee eee ee Le ee ee ee ee ee st YY sm e fe ff mee me Fee J mee fe ei eee men eee me fe ee ef ee me eee ef ef me kand kand band bad bend eee band band ee ee ee ee ee ee ee band kand ee ee baad kad baad kand kaad kaad baand baad band land pa oliocoocciooCcaicHsiqacacHoaacaiqchoaQGeccoaHacoce Ss amp Ves IRS E X Mee ks E HAs E ee E E E E E Ee E hae Hee E E gt i j ke E Heb E E Hae E Ki E E JE E E E i ip Hee pe Ee Figure 35 The Classification Distribution window shows the percent reactivity of each individual spectra You must print this window or manually record the data to be removed as this window will m
50. jects together and is recommended for displaying cell and tissue samples 3 Use the stretch options to obtain the best contrast of objects against the image background Stretching can also reduce the appearance of vertical stripes that are caused by noise in the HSI images Going clockwise see Figure 17 these four stretches show the gold nanoparticles at increasing contrast against background The strength of the stripes also increases The Square Root stretching makes both faint and bright objects more visible Stretching the display does not affect the image data Linear 2 and Equalization stretching show the faintest objects Bene Tools Window T anear Filter r Square Root Image Linear Image Linear 0 255 Image Linear 2 Image Gaussian Image Equalization Image Square Root Zoom Linear Zoom Linear 0 255 Zoom Linear 2 Zoom Gaussian i z Zoom Equalization Linear 2 Zoom Square Root Scroll Linear Scroll Linear 0 255 Scroll Linear 2 Scroll Gaussian Scroll Equalization Scroll Square Root Histogram Matching Interactive Stretching 4 A q Hf a yi Figure 16 Figure 17 Examples of image with faint Enhance drop down menu and bright features using different stretch options 14 Interactive Stretching ENVI also allows the image contrast to be set interactively This mode is selected at the bottom of the Enhance drop down menu Interactive stretching In
51. lick okay in Output Spectral Library window b G d e f g h 4 Output Plots to Spectral Library A Spectral Library Plots y Select Plots to Output Adee 212 450 bands Aldo 11 450 bande 0199 Y 116 450 bands Ace 176 450 Bands 200 7 11 450 bands Aali 178 450 bands A200 119 450 Bands Acad 120 450 Bands A201 7120 450 bands Acad 119 450 bands A200 7 118 450 bands Aral w119 450 bands A202 2120 450 bands Aaz v1 21 450 bands scm Edit Options Plot_Function Help Input Data gt Save Plot As gt ASCII Spectral Library Tra IDL Variable Cancel PostScript Image File Figure 38 Saving Spectral Library Number of items selected 0 Select All tems Clear All tems EJ Output Spectral Library Spectral Library Header Information 2PoiRangel lof XAvieTile Wavelengh vass T le Vaue J Reflectance Scale Factoe 1 00 h 2 fe Wavelength Units Nanometers My Recent 4 KE140 z eia Input to Output Data Scaling Ej laJeF140711 X Seale Factor 1 00 Desktop IKF LA T Scale Factor oo f Output Result to Fie C Memory My Documents Ente Qulput Fdename Choose s My Network File name KFI 40 710_subset_sif hdr Open Places a Figure 40 Saving Output Spectral Library 50 Part 11 Mean Spectra Analysis and Normalization Mean Spectra Analysis and Normalization allows the user to compare spectra from separate samples Th
52. ll see a red border around the image Grab the corner and resize the border to the desired image size then click OK g Click OK in the Select Spatial Subset Window h The Output File Type gives a list of image types supported by ENVI Select one of your choosing 2 To enter the filename and folder a Select Choose b Then enter the File Name c Click Open d Then click OK in the Output Display to Image File 3 The Zoom window can be saved in exactly the same manner a Choosing Save Zoom Image b Selecting Image File oanogo E3 Output Display to Image File Ed Resolution 24 bit Color BSQ Change Graphic Overlay Selections Spatial Subset Full Scene Input Image Resize Factor O00 Output Image Size 320 x2 5x3 Change Image Border Size Output File Type ENVI ha Output Result to File Memory Enter Output Filename Choose C Documents and Settings Owner ODNomalONH Select Mask Clear Mask Figure 26 Output Display to Image File 20 Chapter 3 Extraction of Spectral Profiles The CytoViva Hyperspectral image analysis software powered by ENVI ITT VIS contains several essential features both manual and automated to aid the user in the quantification and identification of materials by spectral data analysis Part 1 Opening Test I mage In this example we illustrate how you can extract spectra from gold nanoparticles that appear in different colors in the RGB display 1 Open the E
53. n Menu Method 2 Using the Endmember Collection Spectra Window Step 3 Performing the SAM Classification Part 3 Re Loading Images from Available Band Lists Window Part 4 The Rule Image Part 5 The Classification SAM Image Part 6 Re Classifying the SAM Part 7 Linking the Displays Part 8 Overlaying the Classification Image Part 9 Merging Colors in a Classification Image Part 10 Removing Unwanted Spectra from Data Analysis Part 11 Mean Spectra Analysis and Normalization Part 12 Resizing an image Part 13 Spectral sub setting an image Chapter 6 Quantitative Spectral Analysis Part 1 Recording of the light source Part 2 Spectral Normalization of the Lamp Part 3 Spectral Normalization of the Specimen Part 4 Correction for second order diffraction Part 5 Conversions to Reflectance and Absorbance Chapter 7 Clean Up and Supplies Part 1 Clean Up of the System Part 2 CytoViva Supplies Glossary Trouble Shooting Chapter 1 CytoViva Hyperspectral I maging System Installation O CytoViva High Resolution Adapter cytoviva Dual Mode Fluorescence DMF Module G VNIR Hyperspectral Imager 4 Optical Microscope G Motorized Stage 6 150W Halogen Light Source optical Camera 8 Dual Port 8 Computer Chapter 2 Quick Start Guide Part 1 Starting CytoViva ENVI Software 1 The CytoViva HSI microscopy system is started by first running ENVI a Double Click the ENVI icon that appears on the desktop to start ENVI
54. nk display window appears see Figure 14 b Open a new HSI image file Press Load to update the blank display with the new image C EEX bT Available Bands List File Options C Gray Scale RGB Color oR Band 449 986 1235 4uNp 40x air MG Band 1 400 6234 AuNp 40x ar D p Band 449 986 1235 AuNp 40 air Dims 696 x 399 Integer BIL Load RGB No Display v New Display Figure 14 New Blank Window Figure 13 Available Bands List Window 2 Multiple HSI images that have already been opened and appear in the Available Bands List can be displayed in separate windows To open multiple images from the Available Bands List see Figure 15 Right click on the desired image Load True Color to lt new gt will load the image into a new window Load True Color to lt current gt will load the image over the current image Load Default RGB to lt new gt will reset RGB values and open a new image oango 9w window File Options p AuNp Dims 696 x 399 Integer BIL Load RGB Display X Available Bands List BEE Open File in ENYI Zoom Add Selected File to Memory Close Selected File Edit Header Quick Stats Load True Color ta lt new gt Load CIR to lt new gt Load True Color to lt current gt Load CIR to lt current gt GiaySea Load Default RGB to lt new gt Load Default RGB to lt current gt Fold 4ll Files m
55. of the particles into the classification 1 2 SA i n en Go back to the Spectral Angle Mapper Parameters window To return to the Spectral Angle Mapper Parameters window click apply in the Endmembers Collection SAM window The Spectral Angle Mapper Parameters dialogue box opens Change the Maximum Angle from 0 1 to 0 3 This time do not create a Rule image Now click Preview The new classification result is shown to the right of the dialog You can see that more particles are classified this time For Output Result select Memory For Output Rule Image select Yes For Output Result select Memory Then click OK From the top of the available bands list right click on the new image Then click on Load Band to New Display The new classification 2 SAM images Figure 22 with the original image Figure 21 and the original classification 1 SAM Image at the side Figure 20 When you compare the first and second classification images you will see that additional particles scattered about the cell are classified as GNR when the angle was set to a greater value The correct value for the angle must be found by experimenting with different values 4 2 Zoom 4x Figure 21 Original image of the GNR in tissue Figure 20 Original Classification Figure 22 Right New Classification 1 SAM image obtained using 2 SAM image obtained using 0 3 0 1 Maximum Angle Maximum Angle 43 Part 7 Linking the Displays
56. ool window 10 In the Interactive Merge Classes window see Figure 32 a Select one coordinate X Y in the Base Class list Note The first is automatically defaulted to red 11 In the Classes to Merge into Base list a Select all other coordinates Note Do not select the coordinate chosen in the Base Class list b Click Okay c The classification file will automatically update to a single color see Figure 33 EJ Interactive Merge Classes Ed Base Class Unclassified 7 1 R Band 200 G Band 128 B Band 53 Ngel MM_60x_1 asa iene oo 25 5 1 Interactive Class Tool File Edit Options Help Active Class Unclassified M297 V 27F 2300 295 ri oeoa Von E kkssev 272 M On on E kss0v 257 M On On BB ksys M On Von 3467 356 M On lt Figure 33 Classification spectra unified to a single color Number of items selected E Select All tems Clear All Items OF Cancel Figure 32 Interactive Merge Classes window 47 Part 10 Removing Unwanted Spectra from Data Analysis This feature provides you with the capability to remove spectra once data analysis is complete and a classification file is created 1 In the original image window select Overlay from the menu bar a Then select Classification b Select the appropriate file 2 The Interactive Class Tool window appears see Figure 34 a Select Options 3 Interactive Class Tool File Edit Options Help Act
57. overwritten by the next scan and the old data will be lost Part 3 Starting the Scan 1 2 After entering the camera settings the HSI scan is started by pressing Collect Hyperspectral Image at the lower left A Live Preview of the spectral image can be included before the scan to help in the selection of the camera exposure time This information is used to a confirm the placement of objects in the field of the scan b to aid in adjusting focus c to confirm that camera exposure is appropriate To see the Preview dialog check Show Live Preview before starting the scan Show Live Preview was selected the HSI Microscopy Preview screen appears see Figure 3 This screen is used by first putting the cursor inside its borders If the cursor is outside the window some information will not be displayed The upper panel of Preview shows a histogram containing the number of image pixels that have values between zero and the maximum value Under the histogram the panel displays a live spectral image that is being captured by the camera In this image the horizontal axis represents the X axis of the HSI image that will be recorded The vertical axis represents the wavelengths of the spectrum captured in the HSI image In the image below the vertical streaks are spectra from objects that are in the view of the HSI camera This image shows that two bright objects are in the view Overlaid onto the spectral image is an intensity plot Placing th
58. owing which pixels exceeded the threshold for matching the endmember 1 Inthe main menu bar select Spectral a Then select Mapping Methods b Then select Spectral Angle Mapper c Then select the file in the list to use as the input image For this exercise select Gold nanorod in Tissue 40Xair img see Figure 10 J Classification Input File File Information File E CTL Test HSI Images Gold nanorod in Tisst Dims 696 x 693 x 468 BIL Size Integer 451 459 008 bytes File Type ENYI Standard Sensor Type Unknown Byte Order Host Intel Projection None Pixel 0x0 Meters Wavelength 400 038513 to 1000 129028 Upper Left Corner 1 1 Description CytoViva HSI Microscopy image Date Thu Jul 17 16 16 26 2008 Camera vendor Unknown CCD dimensions 1392x1024 Exposure time ca Figure 10 Selecting the input file for classification Select Input File Gold nanorod in Tissue 40Xair ima 2 For this specific example wavelengths above 800 nm provide no information for this sample image So to exclude the bands covering wavelengths above 800 nm use the Spectral Subset feature SAM will create a 400 dimensional vector from 400 nm to 800 nm Click on the Spectral Subset button All bands are highlighted by default First remove the selection from all bands by clicking Clear 6 Then in the boxes to the left of Add Range enter 1 in the far left box and 312 in the box E File Spectral Subset Select Bands
59. plot and selecting Plot key see Figure 20 right plot To change the title and axis labels a Right click b Select edit c Then select Plot Parameter d This dialog is used to change scales of the axes the background color style of axis markers and plot margins e The Edit Plot Parameters dialog also allows the right margin of the plot to be made wide so that longer names are displayed properly Change the number in Right Margin to a larger value The label at the right margin shows the coordinates of the pixel used to create the spectrum The label can be changed to a name that refers to the spectral feature By right clicking the window Selecting Edit Then selecting the Edit Data Parameters dialog Select labels appearing in the list and enter a new name in the box below Click on Apply In the plot at the right the labels are changed to refer to different particles see Figure 25 BJ ROI Statistics Results AU100 40Xoil 2 2 250ms E E K f7 RON Statistics Results AU100 40X0il 2 2 250ms O x File Options Stats for Region 1 Red 324 points Stats for Region 1 Red 324 points Select Plot v Clear Plat Select Plot gt Clear Plat ROI Means AU1CO 40Xoil 2 2 250ms i Partical Manow sity Rel Units FOG 600 900 10C ZQ wavelength Select Stat gt Select Stat gt Filename C Documents and Set tings Chuck Ludwi ROI Region 1 Red 324 points ROI Region 1 Red 324 points E Basic Stat
60. r file includes information for CytoViva ENVI to allow the image program to be opened and used Endmember This is any spectra that have been chosen as important for identification purposes in a sample Output File This is a file that you create to include several different output files such as Spectral Library Files Header Files Spectral Angle Mapper Files etc This file will allow you to easily access this information when needed Z Profile Average Window Setting the size windows will either decrease or increase the noise in the plotted spectrum Plot Key This is the key that contains the labels of your spectral curves chosen X Axiz Default The X Axis default is the Wavelength By moving the cursor over the point of the peak you can see the wavelength of the spectra in the selected location Y Axis Default The Y Axis is the Value This provides the intensity levels of the spectra Data Parameters this can be found in the plot key by click on the image menu bar and selecting edit then select Data Parameters The Data Parameters will allow you to change the Key Plot names line style dotted thickness and color Plot Parameters this can be found in the plot key by click on the image menu bar and selecting edit then select Plot Parameters The Plot Parameters will allow you to change the title X amp Y axis background and foreground colors and the font Spatial Resolution The minimum spatial separation that can be d
61. ral Library dialog box open a Select all of the plots listed in the window by clicking the Select All Items Button b Then click OK You will see the Output Spectral Library dialog box open a Click the select all items button to put all spectra into the spectral library Also individual spectra can be collected and saved into the spectral library as well b Then click OK the Output Spectral Library window will open In the Output Spectral Library window enter the baseline intensity and the highest y axis value of the three spectral curves in the Z Plot Range boxes a If curves came from imagery where the dark background had automatically been subtracted enter 20 This value will allow the variation around zero to be included b Do NOT click OK yet This information is needed for scaling spectral library curves to the size of 6 other curves For now the other boxes should not be changed In the Output Spectral Library box click the Choose button located beside the enter output filename to select where the data will be saved a Goto your Output Folder and type in the chosen filename b You do not need to add an extension to the file name c Click Open i The library file and a header file are automatically saved in the Output Folder ii The library is also added to the top of the available bands list d Then click OK 29 Part 3 Viewing the Spectral Library File The Spectral Library Viewer allows you to view th
62. re Cancel Figure 3 Frame out of focus Figure 4 Frame in Focus Camera exposure The intensity plot can be used to adjust the camera exposure Begin by moving the brightest image features to the crosshair If the exposure is too long the intensities will be seen to flat top on the plot see Figure 5 If this should occur the camera is overexposed to these areas and data will be missing in the HSI image If the exposure is too short the spectral data will be noisy The best exposure will be obtained when the intensity of bright features is near 3000 Intensities of a count under 200 don t yield sufficient spectra data Intensities will clip around 4000 The intensity values appear on the plot ordinate 11 12 The plot scale will adjust to more and less intense parts of the spectral image as you move the cursor The most intense readings will occur at the spectral peaks and thus the cursor should be placed at the brightest portion of the spectra to insure that the HSI image will not flat top any of the recorded spectra Usually the spectral peak is near the center wavelength so the cursor would be placed halfway between the lower and upper ends of the spectrum A CytoViva HS Microscopy P Capture Cancel Figure 5 Intensity Plot showing flat top Make changes to the camera exposure by canceling the Preview window and returning to the Camera Settings dialog After making a change go back to the Liv
63. re 27 scroll through the Available Bands List to find the original image Click on the filename to highlight it Click Display and select New Display Click Load Band to display the image In the picture window see Figure 28 a Choose Overlay b Click Classification 5 Available Bands List DAR File Options i Memory3 A o Sam Gold nanorod in Tissue 400 lt air img H Memory2 o Rule Au NR Tissue FE Memory o Spectral Library ff Gold nanorod in Tissue 40Xair ime 3 Band 1 400 0385 Open File in ENVI Zoom 3 Band 2 401 3245 Add Selected File to Memory Band 3 402 6100 1 Band 4 403 8955 Band 5 405 1815 Edit Header a Band 6 406 4670 3 Band 7 407 7525 Quick Stats E a E EI y1 R Band 700 G Band 178 B Band 53 Neecl MM 60x 1 Close Selected File 1 Band 8 409 0385 1 Band 9 410 3240 1 Band 10 411 6095 Bosh 1410 DORR Load True Color to lt current gt rc Gray Scale RGB Color Load CIR to lt current gt Load Default RGB to lt new gt CR Rule Au NF Tissue Memoy2 Load Default RGB to lt current gt ea Emanaa Godna OH Al les Figure 28 Image to be given an overlay Unfold All Files CB Band 48 460 4655 Gold nanorod in Tissue 4 Dims E98 693 Floating Point B50 Load RGB Display 2 Figure 27 Available Bands List Load True Color to lt new gt Load CIR to lt new gt 6 The Interactive Class Tool Input File window appears see Figure 29 7 Sele
64. re 3 Spectral profile of particle with label in plot key 5 33 Step 2 Saving the endmember to the spectral library 2 methods Method 1 Using the Main Menu From the Main Menu bar in the Image Window select Spectral a Then select Spectral Libraries b Then select Spectral Library Builder 2 You will see the Spectral Library Builder dialog box open a Then select First Input Spectrum and click OK In the plot window click and drag the label AU NR Tissue to the spectrum list in the Spectral Library Builder dialog see Figure 5 k cy Spectrum Hame Color 1 SU WA tissue Aled ra a Select Al Pot Delete Figure 5 Spectral Library Builder dialog window 4 Click Plot to create a plot window that is used to accumulate the Endmember Collection Spectra see Figure 6 In this case we have a single endmember EEk E Endmember Collection Spectra File Edit Options Plot Function Help Endmember Collection Spectr oo0 7or e m m r a a a Gi YOU aL Figure 6 Plot containing the spectrum of the endmember6 34 Method 2 Using the Endmember Collection Spectra Window 1 Inthe Endmember Collection Spectra window click on File a Then select Save Plot As b Then select the Spectral Library see Figure 7 2 The Output Plots to Spectral Library dialogue box opens a Click Select All Items b Then click OK 3 The Output Spectral Library dialogue box opens a In t
65. require a longer scan If the area of interest is located only in the center of the field of view near the intersection of the eye piece cross hairs 200 300 lines will be sufficient Default Bands to Load The spectral image bands that will be used to represent red green and blue are selected using the drop down lists for each color The default bands will give a true color representation of the data When the HSI data does not contain intensities at a default wavelengths other image bands can be selected to represent one of the colors To choose a new band to represent a color press the down button A list of the HSI bands and associated wavelengths for each band appears Scroll through the list and select a band The dialog jumps automatically from red to green and to blue as new selections are made Output Files Dark values can automatically be subtracted from the image data during the scan Check the box to enable this feature Enter the number of frames to average for subtraction The default value of ten is recommended The folder and name that will be given to the HSI image is selected next SelectChoose Browse to the desired folder and enter a new file name A new folder for the file can be created if desired It is helpful though not necessary to add the extension img to identify the file as an HSI image mportant _If an old filename is used for example by rescanning without changing the filename the old file will be
66. right of the in the bottom left corner of the zoom window The crosshair cursor appears below in the Zoom window see Figure 2 Figure 2 Zoom window with crosshair enabled and centered on particle 4 Pulling up the plot window a Using the Image Menu Bar i Select Tools li Select Profiles iii Then select Z Profile Spectrum b Or right click in the zoom window and select Z Profile Spectrum 5 Inthe Zoom window move the cursor now in a crosshair format over a red object 6 Left click in the zoom window to obtain the spectra of the pixel under the cursor in the spectral profile window see Figure 3 a The arrow keys on the keyboard will move the crosshair to the desired position Then right click in the plot window and select Set Z Profile Average Window Adjust the two window size boxes to 3 Click OK You will see that there is a decrease in the noise of the plotted spectrum as the spectrum is now averaged over 9 pixels centered on the crosshair oe EJ 1 Spectral Profile Gold nanorod in Tiss mE File Edit Options Plot_Function Help Spectral Protile oe a ae ae ae SIL BO r OL aL UG 106 139 577 46 373 68750 Wavelength Figure 3 Plots using the Spectral Profile tool 22 Part 3 Reading the Data The spectrum contains data from all the image bands at a specific set of pixels chosen in an ROI or a spectral profile The band number wavelength and spectral intensity can be displayed at the bottom l
67. s Min Mean Std 496914 203704 Filename C Documents and SettingsSChuck Ludwi Basic Stats Min Mean Std Band Band 496914 Band Band 203704 Band Band 432099 Band Band 956790 Band Band 083333 Band Band 320988 l an Band Band 021605 Band Band 793210 Band Band ENERE 432099 956790 083333 320988 021605 gt Pelee ee DO fs fmm ef ef ODO Won nron amp woMe O O O ODO OOOO yO 0 n o one wa Nne Figure 25 Spectral plots of ROIs 19 Part 12 Saving Images The hyperspectral image data is saved when the image is scanned before an image is displayed This is why the filename for the HSI file must be given before scanning There is no need to save this data again ENVI can open images from files and display them or bring selected HSI bands into a gray scale or RGB image display These display images can be saved in various formats It is possible to save the main image and the zoom image 1 Inthe main ENVI bar click on File Then select Save Image As Then select Image File The Output Display to Image File dialog opens see Figure 26 If the image is color use 24 bit BSQ resolution For gray scale images use 8 bit A portion of the image can be saved using Spatial Subset The number of samples NS is equal to the number of image columns The number of lines NL is equal to the number of lines scanned and to the number of image rows f Click the Image button to create the portion desired You wi
68. s the zoom while decreases it It is easier to center the cross hair on image features using a higher zoom setting BY 1 Spectral Profile Purdue AU nanorods E B R File Edit Options Plot_Function Help Spectral Profile Remove 466 91 3x3 AYG v Plot Key Stack Plots Set Stack Offset v Collect Spectra Set Z Profile Avg Window v Auto Scale Y Axis Reset Range 50C 600 700 800 900 1000 Wovelen gth Figure 21 Window to collect spectra from multiple regions Noise Reduction Noise Reduction can be applied to the Spectral Plot by averaging pixels Averaging Pixels 1 To reduce noise in the spectrum right click in the Spectral profile 2 Select Z profile average window see Figure 22 3 By increasing the average window size to 3x3 this will significantly reduce the noise on the spectrum Larger window sizes can be used to further reduce noise if the spectrum of interest comes from a feature that is at least this size Excessively large window sizes will begin to include spectra from areas unrelated to the feature changing the spectrum Use smaller window sizes for smaller materials of interest 4 The Z Profile is useful for quickly sampling the spectrum from any part of the image EJ Set Z Profile Aver 2 Window Size aS x 3B OF Cancel Figure 22 Z Profile Average 5 The spectral profile see Figure 23 is marked with red green and blue lines that show the wavelengths of the RGB display
69. splay the names of each spectra collected The default names will be the X and Y locations of the pixel selected for the spectra collected c Then go to edit plot parameters This will allow you to change the title of the plot X and Y axis title d By selecting Data parameters you can change the line color line style and the plot key names e By selecting editing plot parameters you can assign the spectra to different colors f Finally change the names and colors of these new spectra so they can be recognized later Spectral Prafile BOG Oe YOO BOG GoGo 1B0G wavelength Figure 7 Spectral profiles from three image features plotted together 26 Part 7 Deleting Unwanted Spectra from the Plot If you decide to remove a spectral curve this can be done without restarting the plot 1 In the plot window right click on the label of the unwanted spectra collected Note the first spectrum with label at the top of the list cannot be removed 2 Then click on remove label see Figure 8 3 The selected curve will be removed ABE 191 Remove 466 991 3x3 aye w Plot Key Stack Plots Set Stack Offset w Collect Spectra Set Z Profile Avg Window w Auto Scale Y Axis Reset Range my rm Te m eee PaE aa t E a a BOO BOG Foo BOG goo 16 G Wovelength Figure 18 Option list obtained by right clicking on the plot label 27 Chapter 4 Saving and Viewing Profiles in the Spectral Library
70. ssed or plotted together Stats Plots the spectrum of each selected ROI Example of creating an ROL 1 Click on the ROI Type menu 2 Then select the polygon drawing tool 3 Select the zoom window display see Figure 24 4 Outline the boundary of a feature by first clicking near the feature Then drag the cursor around the feature clicking at points where the edge of the ROI changes direction Return to near the first point and right click once 5 After the first click a handle is shown that can be used to move the ROI After the second click the area encircled will fill in with a solid color marking the ROI on the image for examples created with the ellipse and the polygon drawing tools 7 1 R Band 188 G Band 118 B Band 48 AU10 Color Pixels Poly Red 324 14322 Green 469 17465 EIAI Zoom 4x BOA oa 2 F i 3 New Resion Goto Stats Grow Pixel Delete Figure 24 The ROI Tool and examples 18 Part 11 Plotting and labeling spectra from ROI s Regions in the ROI list can be selected for plotting mae ad To select all regions press Select All Individual ROIs are selected by clicking to the left of the name in the list Select the ROIs whose spectra are to be plotted Use the Stat button to create a plot window containing the spectra for each ROI The spectral curves from the red and green ROI s above are plotted see Figure 20 left plot A data legend was added by right clicking in the
71. t to reduce the strength of the recorded spectrum at longer wavelengths The idea of spectral normalization is to compensate for these effects so that spectral characteristics of the sample are clearly seen in the data To perform this it is first necessary to have the light source spectrum that was obtained with your HSI system 56 Part 1 Recording of the Light Source The lamp spectrum has already been recorded on each HSI microscope and provided as a spectral library called Halogen Lamp Spectrum If you have an older system you can obtain the lamp spectrum with the following procedure 1 Sad If you are using a CytoViva condenser a Remove the liquid light guide from the input port by unloosening the screw and pulling out the liquid light guide Set up a scan a Set the exposure time to 0 005 seconds and the number of lines to 20 Keep all other settings the same b Enter a file name like Light Source and start a preview window Setting the light source up a Move the lens turret to an empty position b Aim the end of the liquid light guide into the empty position and watch for the light to register on the preview window c You will want to tape the guide in place while you run the scan Make sure that the values in the Preview data are under 4000 Start the scan After finishing set a rectangle ROI in the main image window Select Stats in the ROI tool The spectrum along with the maximum minimum and standard deviation spe
72. tains will modify the spectrum differently by virtue of their distinct light absorption properties This situation is common for all of the microscope methods that directly sample the illumination such as bright field reflected bright field and dark field modes The situation is different for spectra of nanoparticles made from the noble metals For these recorded spectra show light scattered within narrow peaks that are caused by plasmon resonances where nanoparticle electrons resonate at frequencies that are influenced by the dielectric materials properties surrounding the particle When recorded with the dark field method the recordings are exclusively from the scattered light and do not contain any of the light absorption features So instead of recording the light source spectrum after it has being modified by features of the specimen the recording contains discrete resonance peaks whose strengths are set by intensity of the light source at the resonance wavelengths Of course if the specimen contains dyes pigments and nanoparticles both kinds of effects can occur together in the recorded spectrum For all types of HSI spectra simple procedures can be used to improve accuracy and repeatability so that results can more easily be compared between recordings One of the simplest methods is spectral normalization This procedure is able to counter the effects of the camera and spectrograph on the spectral recording Both of these components ac
73. teractive Stretching with Input amp Output Histograms 1 The display can be adjusted by changing the relationship between the input and output see Figure 18 histograms This option provides greater flexibility to adjust the image display The vertical dashed lines at the left and right edges of the input histogram can be moved to reset the bottom and top of the display range Alternatively these boundaries can be set by typing the new values into the stretch boxes The new stretch is made effective by pressing Apply Image noise is effectively removed from the display by setting the lower boundary value between 5 and 20 For RGB images the process is repeated for each color by clicking the R G or B Apply Stretch D E24 RA GCB Input Histogram utput Histogram HH a a Current Square Root Hist Source Image 160 000 ports 4 A Input histogram of the red image band Output display histogram for the red channel for the red channel Figure 18 Input and Output Histogram 15 Part 9 Plotting Intensity profiles along the X Y and Z spectral axis can be plotted using the profile tool at the top of the main image window 1 Click on Tools then high light Profiles see Figure 19 QDOT tissu 5J 2 R Band 187 G Band 117 B Band 47 Au10 EBR File Overlay Enhance JERE Window Link Krofe Polarization Signatures gt Profile Region OF Interest gt Color Mapping K Z Profile Spectrum
74. the individual lines of the HSI image are acquired until the stage has moved to the end of the scan area During this process the percentage of completion of the scan is reported on the progress bar Ignore the percent value in the lower right corner After the scan finishes the HSI image automatically opens in ENVI iT CytoVive HS Microscopy Acquinng Hyperspectral Image Output File C Documents and Settings Chuck LudwighDesktops 4 lil Cancel Figure 8 CytoViva HSI Acquisition Window Important if an image was rescanned using the same filename as a previous scan the old data is lost However the display still contains the old images If this happens remove the old data from the Available bands list by selecting close all files and open the newly scanned image into the List 16 17 18 To remove the old data highlight the old data in the List Press File in the main image and click on Close Selected Files Click through the display warning and the old data will be removed Open the newly scanned File to place it in the List and obtain the new image See Open an Existing File below for more details Part 4 Displaying HSI Images The HSI image display includes three windows Image This window is sized to include a user selected area of the full HSI image The default is a square window of 400 x 400 pixels The image can be resized by pulling at the lower right corner with the cursor S
75. tions E Memory A O Sam Gold nanorod in Tissue 40air img Memoarny2 O Rule Au NA Tissue Mernory O Spectral Library E Gold nanorod in Tissue 40air imou o Band 1 400 0385 pen File in ENYI Zoom Band 2 401 3245 dd Selected File to Memory Band 3 402 6100 Band 4 403 8955 ut Band 5 405 1815 Edit Header Band 6 406 4670 Band 7 407 7525 ce eek Band 8 409 0385 Band 9 410 3240 Band 10 411 6095 Br Te Bri POR Load True Color to lt current gt C Git Seats RGE Color Load CIR to Saes Close Selected File O BAA sa 0 Load True Color to lt new gt Load CIR to lt new gt AHA AR J i Load Default RGB to lt new gt Cop Rule Au NR Tissue Memory2 Load Default RGB to current l Fold All Files Band 118 550 4625 Gold nani Unfold Al Files cop Band 48 460 4655 Gold nanorod in Tissue Dims 696 x 693 Floating Point BSQ Load RGE Display 2 Figure 17 Available Bands List 40 Part 4 The Rule I mage The Rule image will show those areas that best match the Endmember Spectrum as being darker Areas with a poorer match are lighter For the cell cluster the cell areas are gray and the background around cells is white The GNR particles will appear darker than the tissue The area around particles will also be darker indicating that a zone around the small particles also contains the signature for that particle These zones are produ
76. ue 40air img Band 301 785 7415 Gold nanorod in Ti ssue 40Xair img Band 302 787 0275 Gold nanorod in Tis Hair img Band 303 788 3135 Gold nanorod in Tissue 40air img Band 304 789 5985 Gold nanorod in Tissue 40 air img Band 305 790 8845 Gold nanorod in Tissue 40 lt air img Band 306 792 1705 Gold nanorod in Tissue 40 air img Band 307 793 4554 Gold nanorod in Tissue 40Xair img Band 308 794 7415 Gold nanorod in Tissue 40air img Band 309 796 0275 Gold nanorod in Tissue 40Xair img 30 Gold nanorod in Tissue 40air img Band 311 798 5985 Gold nanorod in Tissue 40air img Band 312 799 8845 Gold nanorod in Tissue 40 air im Band 313 801 1700 Gold nanorod in Tissue 40air img 554 Gold nanorod in Tissue 40 air ima Number of items selected 31 2 Add Range Select All Clear Import ASCII OK Cance Previous Figure 12 Selection of a subset of the spectral bands 10 You will see the Endmember Collection SAM dialog box open 11 From the Plot Window drag the label AU NR Tissue to the Endmember Collection SAM box It is highlighted automatically in the list see Figure 13 12 Click on Apply gt EnimembenCollection SAM n x File Gold nanorod in Tissue 4isar img Full Scene Mask none selected Wavelength 400 0385 to 1000 129 unknown unit 468 bands 1 468 Spectrum Name AuNR Tissue ill Select t Plot Delete Apply Cancel Figure 13 Endmember Colle
77. um at the peak of this spectrum where also the light reflectance is minimum Reflected Light To obtain absorption spectrum of an opaque sample material use a reflected light illumination which sends light that bounces off the sample back into the microscope objective Sg dh ae ee Locate the area of interest on the specimen and center it in the scan area Draw an ROI around the region Save the mean spectrum in a plot After recording the specimen replace it with a reflectance target A suitable target is made by LabSphere model SDM 100 DU or SDM 050 DU Open the ROI tool and use the same region to obtain the mean spectrum of the region Save this to a plot Calculate the percent reflectance spectrum using the spectral math expression a 100 float S1 S2 Assign S1 to the spectrum of the specimen and S2 to the target spectrum Choose a new window to create the absorption spectrum The light reflectance is maximum at the peak of this spectrum where also the light absorption is minimum 61 Chapter 7 Clean up and Supplies To properly take care of the CytoViva Hyperspectral system there are a few steps that are necessary to do when you are done using the system to keep the system performing at it top level Part 1 Daily Clean Up of the System When you are done using the CytoViva Hyperspectral Imaging System it is important to clean the excess immersion oil off the condenser and the objectives Also turning off components of th

Hyperspectral User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents