LabMicrobe manual
Contents
1. ccceccecseseeeeeeeeeeeeeeseeeeeseeeeeas 16 12 1 Training of the neural network classifier cccccceeeeseeeneeeeeeeeneeeneees 16 i 6 ae GF Sto ne 19 DC SUING ee orc sa cseeere E E 21 15 The File menu Saving and loading settings ccceecesseeeeeeeeeeeeeseeees 22 16 RETELFENCES ccccccccccecccccccecccceccccececcecceueaceceacecauecceneauenesueceensaueeaueneeneas 23 BioRAS LabMicrobe User Manual page 3 of 23 Rapid Analysis Systems for Biology and Life Sciences 1 LabMicrobe general information LabMicrobe is a specialized image analysis environment designed to fully automate the processing of images of microbes or cells It is tuned to analyze high quality epi fluorescent images of bacteria but its open architecture allows easy adaptation to other similar applications LabMicrobe has interactive facilities for viewing the results of the different stages of image processing for experimenting with parameters and for training the internal neural network classifier Classification of shapes is based on contour The program works on 8 or 16 bit grey scale images LabMicrobe is implemented in LabSequencer an application platform for sequential execution of program modules With LabSequencer image analysis can be performed in single steps as a sequence of steps or as a loop for automatic analysis of multiple images LabSequencer allows the user to view every step of the analysis anytime during an2. Tool bar To the left of the image window there is a small toolbar the contents of which depends on the sequence step The following tools are available fe Zoom tab for zooming in or zooming out shift select i Cursor tab dh Move tab for moving the image in the image frame Line tab for measuring distances in pixels for calibration Particle selection tool only available in the Species step 5 Select the sequence execution mode Choose the sequence execution mode in the menu to the lower left side r zj w Run seguence Step Loop Rapid Analysis Systems for Biology and Life Sciences BioRAS LabMicrobe User Manual page 5 of 23 After running a sequence all steps can be viewed and adjusted by simply clicking on the relevant step in the Sequence frame Choose between three execution modes 1 Step goes through the sequence step by step This mode is useful for adjusting image analysis parameters Click on the Go button for each step 2 Run Sequence goes through all steps of the sequence one time This mode is useful for analysis of single images after adjustment of the image analysis parameters Start by clicking on the Go button 3 Loop a directory of files will be analyzed and results saved as individual files for each image or in one file as defined in Results This mode is useful for analysis of multiple images requiring the same program settings 4 The settings
3. 4 images micrometer tif 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq G LabSequencer main vi B an fs LabMicrobe py oleae Detect eger reo 2 Ad rages reserved ne p Pwde an i i Read the line length by clicking on it r Line length in pixels will show toon 106 L Otimege gt UR SDC up along with vesns areni the line angle Set the pixel aerel 2 180 00 0 00 Tc Doane anc anga TT Settings Tere Lato ot a Figure 13 Calibration of particle analysis can be done by loading an image of a micrometer slidetaken with the same microscope and magnification as the images for analysis A Run Lab Microbe to the Particle analysis step Measure the pixel size by drawing a line between two bars Read the length of the line by clicking on it 180 pixels in this example B Calculate the pixel size Pixel size length between bars number of pixels BioRAS LabMicrobe User Manual page 16 of 23 Rapid Analysis Systems for Biology and Life Sciences 12 Species The neural network classifier LabMicrobe includes a neural network which enables it to classify organisms based on their contour In this step the classification can be trained A newly trained classification is saved with Settings just like all other parameter settings 12 1 Training of the neural network classifier Training of the neural network classifier is done by simply pointing and clic
4. Figure 10 Epifluorescens microscope image of a stained bacterial sample loaded into Lab Microbe i LabSequencer main vi Sequence Load image file La b M cro be aap 2006 Detect edges version 2 1 All rights reserved Threshold Morphology Dividing Particle analysis Species Classify Results 1184 328 3 Step 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 11 Dividing cells are marked 1 non dividing cells are marked 0 Rapid Analysis Systems for Biology and Life Sciences BioRAS LabMicrobe User Manual page 13 of 23 Table 1 Some examples on non dividing and dividing cells from a bacterial sample 0 indicates non dividing cell 1 indicates dividing cell Top row shows crop of original images bottom row shows images from Dividing sequence Non dividing Non dividing Dividing Dividing 11 Particle analysis and unit calibration In this step each particle is analyzed and given an individual number Fig 12 The pixel size is used for calibration and should be set to the length of a pixel in real world units lf you do not know the pixel size you can measure the image of a scale bar using the line tool which tells you the length of the line in pixels see Fig 12 For example if you measure a micrometer scale showing 10 um and the number of pixels is 180 the pixel size is 10 um 180 0 055 um pixel or 55 nm pixel since numbers have to be gt 1 Writ
5. Morphology Threshold Dividing as a lt a 5196 Partide analysis my ak sch Mm ae ee aa H a l 4000 species is 7 2 z Classify Re 8 i ait Results pa eae E s i arth 0 2000 4000 6000 Change threshold a See if particles are E O DERE E dark Particles should be green 00x 1200 1 2 16 bitimage 0 238 458 with a green halo around H The default values are positive numbers which should be changed step to negative numbers Invert the default threshold values for dark 4 C Documents and Settings Wick Local setings Teme LabMicr abe sq j p a rt C eS ag ai n st g ht b ac kg roun d OBS The lower number in the lower 5 box Simply write the new numbers in A er the number boxes a Figure 5 Changing of threshold values In this image threshold values should be changed since particles are dark With correct threshold setting particles are green surrounded by a green halo see Fig 6 BioRAS LabMicrobe User Manual page 9 of 23 Rapid Analysis Systems for Biology and Life Sciences le LabSequencer main vi Sequence i abSequencer version 1 Load image file La b M IC robe e aE vee Detect edges version 2 1 All rights reserved Threshold Dividing Co ee eet i a a n ee Partide analysis Species res By ce en Classify ate a B n t 2000 Results tat oe Oe eee z 5195 4000 Particles should be green with a green halo around like shown in this exa
6. appended to one common file Individual files will have the same name as the image they are a result of with the extension txt le LabSequencer main vi Sequence lie LabSequencer version 1 0 Load image fle LabMicrobe ras 2006 Detect edges version 2 1 All rights reserved Threshold Morphology Table Histogram File Dividing Particle analysis C Documents and Settings Nick My Documents LabSequencer Classify 3 LabMicrobe manual Choose a file directory for saving result from LabMicrobe analysis Destination One file for each image Append data to LabMicrobe tet 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 21 Export of data from LabMicrobe The following data are available Index of cell particle Cell type defined by neural network classifier Area of each cell particle Indication of whether the cell is dividing or not Volume of cell particle assuming rod shape capped by two half spheres O O O O 0 A number of potentially useful parameters are calculated in addition to these but they are currently not exported Contact Bioras to request additional parameters 15 The File menu Saving and loading settings For saving settings and Training sets go to the File menu and click Save settings Table 2 For loading saved settings click Load settings in the file menu For returning to the default settings click Default setti
7. can be saved see below and used for analysis of other similar images With LabMicrobe any number of settings can be created 6 Image files and formats The first step in the sequence of program steps is the loading of image files The program contains a sample image of fluorescent beads against a dark background Fig 1 that will appear when the program is started for the first time or if you choose to revert to default settings Select a file or the directory of files for analysis from you hard drive The best format for analysis are the non loss image formats TIFF PNG BMP LabMicrobe can read and analyze compressed file formats such as JPEG The result of the analysis will depend on image quality To load images 1 Click the folder icon M to the right of the file path 2 A dialog window will appear asking Choose an image file or directory containing image files 3 For analysis of a single image choose the image file on your computer in the dialog window and click the Save button 4 For loading directories of files choose the file directory on your computer in the dialog window and click the Select Cur Dir button 5 The image or image directory has been loaded and is ready for analysis 7 Detect edges In this step LabMicrobe automatically detects edges by a Marr Hildreth operation The Marr Hildreth algorithm is a method for detecting edges in digital images where there are strong and rapid varia
8. same type can be selected and added to the example count 8 Repeat the procedure for the next class of particles organisms This will enable LabMicrobe to recognize more than one type of particles Examples of classification of marine bacteria according to shape are shown in Figure 15 9 If the sample contains unwanted particles that should not be counted remember to train Lab Microbe with these by creating an Unknown or 2 type of particles 10 When a suitable number of particles have been chosen for each Train particle type around 10 20 of each type click the L button Higher number of descriptors renders a better training result see Fig 16 11 A training overview will be shown for a few seconds showing targets and outputs Fig 17 Ideally the output should de aligned with the target 12 Training can be continued by adding more selections at any time This can be repeated as many times as necessary 13 One or several Training sets can be saved see below Clear all 14 Start a new training set by clicking to remove old training sets BioRAS LabMicrobe User Manual page 18 of 23 Rapid Analysis Systems for Biology and Life Sciences i LabSequencer main vi m Sequence li abSequencer version 1 Load image file La b M IC robe pesca ee 1 0 a poche All rights reserved Threshold Morphology Dividing Partide analysis Example count D Classify Clear all Add selectio
9. B O R A S LabMicrobe User Manual page 1 Rapid Analysis Systems for Biology and Life Sciences LabMicrobe Description and User Manual Version 2 0 July 2010 Rapid Analysis Systems for Biology and Life Sciences BioRAS LabMicrobe User Manual page 2 of 23 LabMicrobe Image analysis environment Table of contents 1 LabMicrobe general INMOFIMAUO Me cexzcdveccnasdvssemsiearesetdsvsevaddcesexeldaecesesdveseapsdcias 3 2 Installation and startup of LabMicrobe ccc ceccceeeeceeeeeeeeeeeeeeseeeesaeeesaeees 4 SOONG INC 0 exec ctenends accuetasccavecesenettaqeandeetanctceetcs cede aiazatdeeteretceeee auauetazetauetcsate 4 4 Tool eee 4 5 Select the Sequence execution mode ccceecccceececeeeeeceeeeeceeeeeseeseeseaeeenes 4 6 Image files and TOMMANUS wrist cessaticserntdinesentevrepenersestertior ce euroiroreracena 5 7 Detect CAGES ccccccccecccaeeeccececseeeceucecsuceceucecsucessueessueecsueessusessusessusessasensaees 5 oa 10119 0 6 Denne ee ee ee ee eee ees 6 8 1 Threshold for light particles against a dark background c068 6 8 2 Threshold for dark particles against a light background 008 7 Di IWOMDINOIODY aea e A sd 9 10 Dividing Frequency of dividing cells cceeeeeeeeee eee eeeeeeeeeaeeeeeaeeeeees 11 11 Particle analysis and Unit Calibration ccccecccesceceeeeceeeeceeeeceeeesaeeesees 13 12 Species The neural network ClaSSIfi V
10. If classifications are not correct go back to the training step and add more selections BioRAS LabMicrobe User Manual page 20 of 23 Rapid Analysis Systems for Biology and Life Sciences Below classification of the bead image and classification of a bacterial sample is shown Figure 17 Sequence A 1 LabSequencer version 1 0 Load image file La b M ICro e a Bioras 2006 Detect edges version 2 1 All rights reserved Threshold Morphology D Dividing EE Fartide analysis Peet Species Classify Results Cocci j Step Go f 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq lined tat LabSequencer main vi Detect edges version 2 1 Threshold Morphology Dividing Particle analysis All rights reserved Sequence B LabSequencer version 1 0 Load image file La b M IC ro e Bioras 2006 Species Classify Results A od Coeci Rod Fod 1600x1200 1 2 16 bitimage 0 1118 888 m Aste aa a 4 C Documents and Settings Wick Local Settings Temp LabMicrobe Isq Figure 19 Classification by the neural network classifier for the default bead image A and a bacterial sample B BioRAS Rapid Analysis Systems for Biology and Life Sciences 14 Results LabMicrobe User Manual page 21 of 23 Results are displayed in table format or as histogram Fig 20 Raw data is exported in a format which can be read by Excel o
11. bitimage 0 4 456 E o ste co halal _ T C Documents and Settings Wick Local Settings Temp LabMicrobe Isg Figure 7 The Morphology step The number of Erosions defines the level of removal of small particles in the image Sequence Load image file La b M icrobe ee 2006 Detect edges version 2 1 All rights reserved Threshold Morphology Dividing Particle analysis my Species Classify Results The number of erosions was set to 1 which is too low for this example T C Documents and Settings Wick Local Settings Temp LabMicrobe sq Figure 8 The same image with the number of erosions 1 which is too low for this image The result shows many extra unwanted particles BioRAS LabMicrobe User Manual page 11 of 23 Rapid Analysis Systems for Biology and Life Sciences E LabSequencer main vi Sequence s LabSequencer version 1 0 Load image file La b M IC robe Bioras 2006 Detect edges version 2 1 All rights reserved Threshold Piasco Morphology Dividing Fartide analysis my Species Classify Results The number of erosions was set to 3 which is too high for this example 1600x1200 1 2 16 bitimage 0 662 100 al 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 9 The number of erosions 3 which was too high for this image Countable particles were removed by this operation compare to Fig 8 amp 9 10 Dividing Frequency of
12. d after the analysis Figure 1 shows the main functions of the front panel window es LabSequencer main vi m Help LabSequencer version 1 0 Bioras 2006 All rights reserved LabMicrobe version 2 1 Load image file 1600x1200 1 2 16 bit image 32705 656 188 Source file s 7A C Documents and Settings Nick My Documents My Pictures Spot x 4 images 0 516 beads tif Run sequence _ 9 i i T 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq 1 0 A Figure 1 The LabMicrobe front panel showing an epi fluorescent test image of beads with known size 0 516 um 1 File menu save and load custom settings print and exit program 2 Help menu 3 Sequence of program modules Modules are disabled until they contain valid data 4 Execution mode choose between 3 alternative execution modes Step 1 task at the time Run sequence all program steps in sequence Loop automatic analysis of multiple images 5 Tool bar for the image window for zooming moving or simple drawing In the Classification step more tools will be available see below 6 Image display 7 Image size information 8 A path with a source image file or a directory of image files for batch mode For Rapid Analysis Systems for Biology and Life Sciences BioRAS LabMicrobe User Manual page 4 of 23 processing directories set mode to Loop 9 Revert button Revert to the settings in the las
13. dividing cells LabMicrobe can estimate bacterial growth by the frequency of dividing cells FDC see Fig 10 11 Lab Microbe defines dividing cells as those containing two intensity maxima Table 1 Frequency of dividing cells FDC can be used as an indication of bacterial growth rate In order to calculate absolute growth rates from FDC we recomend calibration of the method with laboratory growth experiments The FDC method was suggested by Hagstr m et al 1979 as an alternative to radioactive tracer methods for estimations of bacterial growth When the method was developed dividing cells were counted manually and good correlation between bacterial growth and FDC was found for natural marine bacteria Newell amp Christian 1981 Good correlations between bacterial growth and FDC were found in samples of marine bacteria counted with LabMicrobe Blackburn et al 1998 BioRAS LabMicrobe User Manual page 12 of 23 Rapid Analysis Systems for Biology and Life Sciences i LabSequencer main vi m Sequence 7 LabGenvencer version 1 Load image file La b M IC robe agea es 1 0 a ela All rights reserved Threshold Morphology Dividing Partide analysis Species Classify Results 1600x1200 1 2 16 bit image 23233 1184 460 Source file s C Documents and Settings Wick My Documents My Pictures Spot J images Image tif A Ste G ge 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq
14. e this number into the pixel size field see Fig 12 B All results will be given in these untis The measurement box shows the values of each particle for a given measurement see Fig 11 A number of factors can be shown choose between the factors in the measurement o Particle number Area Perimeter Length Elongation Hemsiphererod Oo0O00 O BioRAS LabMicrobe User Manual page 14 of 23 Rapid Analysis Systems for Biology and Life Sciences LabSequencer main vi m Sequence 7 LabGenvencer version 1 Load image file 7 La b M IC robe T 1 0 searing on All rights reserved Threshold Morphology Dividing Partide analysis Species Classify Results 1600x1200 1 2 16 bitimage 0 552 404 Pixel size 1 Measurement Area il Measurement box AV Ste G gst here showing Area 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 12 The particle analysis step BioRAS LabMicrobe User Manual page 15 of 23 Rapid Analysis Systems for Biology and Life Sciences ey LabSequencer main vi Sequence A lie LabSequencer version 1 0 Load image file La b M IC robe Bioras 2006 All rights reserved Detect edges version 2 1 Threshold Morphology d Dividing Partide analysis i Species Classify Results j a 1600x1200 1 4 16 bitimage 32767 1452 936 Source file s C Documents and Settings Wick My Documents My Pictures Spot
15. king on a number of organisms from each class LabSequencer main vi Sequence a SREE Se L Load image file La b M ICro be age thie icra LO Detect edges version 2 1 All rights reserved Threshold Morphology Dividing Example count Particle analysis 0 Classify Results Selection tool Name field for filling in names hal 4 C Documents and Settings Wick Local Settings Temp LabMicrobe Isg Figure 14 The Species sequence step with the test image epi fluorescent test image of beads with known size Note the extra tool in the toolbar for selecting particles Training steps 1 Fill in the name field see Fig 14 with the name for the class or species of particles that will be chosen first 2 Click on the selection tool see Fig 14 3 Drag a square around a particle 0 4 When moving 5 a cross Maz The cross marks the particle make sure that it crosses it The square can be resized by dragging the corners the cursor over the square the center will be marked by 5 Select multiple particles by holding down the control button Rapid Analysis Systems for Biology and Life Sciences BioRAS LabMicrobe User Manual page 17 of 23 6 During particle selection zoom in PP and out shift can be used Add selections 7 Click on the L button The selected particles will be added to the Example count After adding one selection more particles of the
16. m ple i00x1200 1 2 16 bitimage 0 412 462 fi 1 Fi Run sequence i Revert 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 6 Correct threshold setting for dark particles against a light background In this example the lower number is 3000 and the higher number 2 9 Morphology In the Morphology step unwanted small particles in the image are removed by erosion Each level of erosion removes particles with width in pixels equivalent to twice the erosion level For example a level of 2 removes particles that are less than or equivalent to 4 pixel widths The level should be set to remove all but the smallest particles of interest Fig 8 9 10 Setting the number of erosions to high might result in removal of countable particles while analysis with a too low number might result in a to high particle count The number of erosions is by default set to 2 The number should be varied between 1 and 3 in order to find the number of erosions that works best for removing unwanted noise in the image for each type of images BioRAS LabMicrobe User Manual page 10 of 23 Rapid Analysis Systems for Biology and Life Sciences Threshold Sequence iF LabSequencer version 1 0 Load image file La b M IC ro be Bioras 2006 Morphology Dividing Particle analysis Species Classify Results Set number of erosions here Typically between 1 and 3 1600x1200 1 2 16
17. ngs The cell radius r and volume V are calculated from Area A length defined as the longest chord I r vf l A t 4 17 4 V 47 r 3 m r l 2r BioRAS Rapid Analysis Systems for Biology and Life Sciences LabMicrobe User Manual page 23 of 23 LabMicrobe always remembers the last setting and opens with that Table 2 The file menu File Help Save settings Ctl S Load settings Ctl 0 Default settings Page Setup Print Window Exit 16 References Saves settings and training sets Loads the saved settings Goes back to the default settings Page setup for printing the page Print any sequence Exit LabMicrobe Blackburn N Hagstrom A Wikner J Hansson R C Bj rnsen P K 1998 Rapid determination of bacterial abundance biovolume morphology and growth by neural network based image analysis Applied and Environmental Microbiology 64 9 3246 3255 Hagstrom A U Larsson P Horstedt and S Normark 1979 Frequency of dividing cells a new approach to the determination of bacterial growth rates in aquatic environments Appl Environ Microbiol 37 805 812 Newell SY Christian RR 1981 Frequency of Dividing Cells as an Estimator of Bacterial Productivity Appl Environ Microbiol 1981 Jul 42 1 23 31
18. ns a5 Species Cocd AN ste G gj P 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 15 Selection of multiple particles by holding down the control button for training the neural network classifier i Figure 16 Example of a aC pe I DOE u oe training set used for the neural yy oA SN we tw tht ra bad Ije A gn gi ad o see t ee J network classifier Images w ry 2 were taken of filtered and Stained sea water samples Ss ate 9 7 0 from the Baltic Sea The x a ma gt ty a objects represent the main fe T oe T k classes of bacteria in this sea A Qi Y NG f area R1 3 rods V vibrio X K sa VP 4 rejected objects From e a r wt T Blackburn et at 1998 BioRAS LabMicrobe User Manual page 19 of 23 Rapid Analysis Systems for Biology and Life Sciences Rodi Vibrio Rejecter Number of descriptors fe Figure 17 The contours of objects are defined using Fourier coefficients The figure shows the levels of details achieved by using different numbers of Fourier coefficients solid lines From Blackburn et at 1998 48 Target Figure 18 The training overview window shows up for a few seconds after pressing the Train button Ideally Output should be aligned with Targets 13 Classify The Classify step shows the classification of particles organisms according to the neural network classifier
19. r other spreadsheets Fig 21 Histograms are only shown internally in LabMicrobe i LabSequencer main vi Sequence Load image file Detect edges Threshold Morphology Dividing Partide analysis Species Classify i Results LabSequencer version 1 0 Bioras 2006 All rights reserved LabMicrobe version 2 1 Table Histogram File Po fem sa a aer fe fees lioz feaa i534 eroxe47_ 7 oc s jise lss fiaa eoviee7 l fo Jo elize soaa a5 e6637 i0 fo D CS seri 949 i51 e477 J12 feo o aie s her poen e ea o 1m7 ferso 1534 6 605647 14 coca File Help LabMicrobe version 2 1 Load image file Detect edges Threshold Morphology Dividing Partide analysis Species Classify i Results Table Histogram Fie o SoS l 150000 200000 250000 300000 350000 Measurement 3 Area 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq LabSequencer version 1 0 Bioras 2006 All rights reserved y 400000 450 000 a iaje ajj Figure 20 Results displayed as table or histogram showing particle area BioRAS LabMicrobe User Manual page 22 of 23 Rapid Analysis Systems for Biology and Life Sciences The data can be exported by clicking the File tab in the Results sequence Choose a file directory on your hard disk where results from analysis should be saved If multiple images will be analysed choose if the files should be saved individually or
20. t saved settings file 10 Busy light 11 Run Execute button 2 Installation and startup of LabMicrobe LabMicrobe can be downloaded directly from www bioras com In case of slow internet connections request a program CD from info bioras com The downloaded version of LabMicrobe is packaged into a zip file Installation 1 Unzip the downloaded LabMicrobe installation file 2 Double click Install Lab Microbe file and the executable program will be installed in the Programs folder The Install LabMicrobe program is a Windows Installer which installs the program and other components If you have a previous version of LabMicrobe installed the installer will actually uninstall the previous version run the installer again to install the new version 3 The first time LabMicrobe is run you will be asked to enter a license key You can acquire a key from info bioras com The key will be a permanent license if you have purchased it from Bioras or you can request a trial key which will allows you to test LabTrack with its full functionality for a limited period of time 3 On line help Most controls have tip strips associated with them These are pop up text Strips that appear when the cursor is moved over them Right clicking on a control also offers a Description and tip dialog that describes the function of the control Ctl h opens a help window which shows the documentation of the control under the cursor 4
21. tions in image brightness The Marr Hildreth edge detection method is simple and operates BioRAS LabMicrobe User Manual page 6 of 23 Rapid Analysis Systems for Biology and Life Sciences by convolving the image with Laplacian and Gaussian operators or as a fast approximation by Difference of Gaussians Then zero crossings are detected in the filtered result to obtain the edges Figure 2 shows the result of applying the edge detector to bead image i LabSequencer main vi m Sequence lie LabSequencer version 1 0 toad mage fie LabMicrobe version 2 1 All rights reserved Threshold Morphology Dividing Particle analysis ny Species Classify Results 1600x1200 1 2 16 bitimage 0 122 446 Run sequence 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 2 Detect edges sequence step 8 Threshold In the Threshold step the threshold should be set to isolate particles from the background In theory the lower threshold should be Zero after applying the edge detector but in practice it can be an advantage to set it slightly higher to remove noise The setting should be constant for a given camera and light source Lab Microbe sets the threshold control scale to the smallest and largest grey scale pixel value in the image 8 1 Threshold for light particles against a dark background For light particles against a dark background threshold values will be positive n
22. umbers The automatic generated thresholds are positive numbers BioRAS LabMicrobe User Manual page 7 of 23 Rapid Analysis Systems for Biology and Life Sciences es LabSequencer main vi Sequence lie LabSequencer version 1 0 Load image file La b M IC robe Bioras 2006 Detect edges version 2 1 All rights reserved Threshold Dividing DEE ae eee eee Atipe E j a806 Fartide analysis my Tah Mw Bee EA cra a z Species Seber na EER LOT ene lee z Classify eee Da ee E ea 6000 Results a Pa 4000 2000 0 2000 4000 5196 F A2 1600x1200 1 2 16 bitimage0 656 360 4 ry Fi Run sequence 4 C Documents and Settings Nick Local Settings Temp LabMicrabe sq Figure 3 Threshold sequence step A lower value of 2 removes large areas of induced noise leaving the particles o interest isolated The upper values should be equal or greater to the scale maximum 8805 8 2 Threshold for dark particles against a light background For analysis of dark particles against a light background threshold values will be negative Fig 6 Figure 4 Crop of image with dark beads against a light background analyzed in the example below BioRAS LabMicrobe User Manual page 8 of 23 Rapid Analysis Systems for Biology and Life Sciences le LabSequencer main vi Sequence iF LabSequencer version 1 0 Load image file La b M IC ro be Bioras 2006 Detect edges Version ee 1 All rights reserved Threshold
Download Pdf Manuals
Related Search