Computational Biology, Part 23 Automated Interpretation of Subcellular Patterns in Microscope ImagesInitial GoalPreliminariesAcquisition considerationsSlide 5Slide 6Slide 7Annotation considerationsPreprocessingPreprocessing (continued)Feature levels and granularityCell SegmentationSingle cell segmentation approachesVoronoi diagramVoronoi Segmentation ProcessPowerPoint PresentationSlide 17Slide 18Slide 19Slide 20Watershed SegmentationSlide 22Seeded Watershed SegmentationSlide 24Feature ExtractionThresholding2D Features Morphological FeaturesSuitability of Morphological Features for ClassificationSlide 29Slide 30Illustration – Skeleton2D Features Edge Features2D Features Hull FeaturesZernike Moment FeaturesHaralick Texture FeaturesSlide 36Pixel Resolution and Gray LevelsWavelet Transformation - 1D2D Wavelets - intuitionSlide 40Slide 41Daubechies D4 decomposition2D Features Wavelet Feature CalculationWaveletsGabor FunctionGabor Feature CalculationComputational Biology, Part 23Automated Interpretation of Subcellular Patterns in Microscope ImagesComputational Biology, Part 23Automated Interpretation of Subcellular Patterns in Microscope ImagesRobert F. MurphyRobert F. MurphyCopyright Copyright 1996, 1999, 2000-2007. 1996, 1999, 2000-2007.All rights reserved.All rights reserved.This is a micro-tubule patternAssign proteins to major subcellular structures using fluorescent microscopyInitial GoalInitial GoalPreliminariesPreliminariesAcquisition considerationsAcquisition considerationsResolution defined as ability to distinguish two “point-Resolution defined as ability to distinguish two “point-sources”sources”Maximal resolution in x-y plane given by Rayleigh (or Maximal resolution in x-y plane given by Rayleigh (or Abbe) limitAbbe) limit1.221.22/2NA/2NAwhere where is wavelength of emitted light and NA is the is wavelength of emitted light and NA is the numerical aperture of the objective; 244 nm for 520 nm numerical aperture of the objective; 244 nm for 520 nm light and 1.3 NAlight and 1.3 NASampling theorem (Nyquist) says maximum information Sampling theorem (Nyquist) says maximum information can be obtained if we sample at twice the maximum can be obtained if we sample at twice the maximum frequency present in a samplefrequency present in a sampleTry to achieve Nyquist Sampling at Rayleigh limitTry to achieve Nyquist Sampling at Rayleigh limitAcquisition considerationsAcquisition considerationsMaintain low cell density if single cell Maintain low cell density if single cell measurements desiredmeasurements desiredControl acquisition variablesControl acquisition variablesSelect (initial) focal plane consistentlySelect (initial) focal plane consistentlySelect fields consistently (at least one full cell per Select fields consistently (at least one full cell per field)field)Maintain constant camera gain, exposure time, Maintain constant camera gain, exposure time, number of slicesnumber of slicesSelect interphase cells or ensure sampling of cell Select interphase cells or ensure sampling of cell cyclecycleAcquisition considerationsAcquisition considerationsCollect sufficient images per conditionCollect sufficient images per conditionFor classifier training or set comparison, more than For classifier training or set comparison, more than number of featuresnumber of featuresFor classification or clustering, base on confidence For classification or clustering, base on confidence level desiredlevel desiredCollect reference images if possible (DNA, Collect reference images if possible (DNA, membrane)membrane)Acquisition considerationsAcquisition considerationsCollect sufficient images per conditionCollect sufficient images per conditionFor classifier training or set comparison, more than For classifier training or set comparison, more than number of featuresnumber of featuresFor classification or clustering, base on confidence For classification or clustering, base on confidence level desiredlevel desiredCollect reference images if possible (DNA, Collect reference images if possible (DNA, membrane)membrane)Annotation considerationsAnnotation considerationsMaintain adequate records of all experimental Maintain adequate records of all experimental settingssettingsOrganize images by cell type/probe/conditionOrganize images by cell type/probe/conditionPreprocessingPreprocessingCorrection for/Removal of camera defectsCorrection for/Removal of camera defectsBackground correctionBackground correctionAutofluorescence correctionAutofluorescence correctionIllumination correctionIllumination correctionDeconvolutionDeconvolutionPreprocessing (continued)Preprocessing (continued)RegistrationRegistrationNot critical if only using DNA or membrane Not critical if only using DNA or membrane referencesreferencesIntensity scaling (constant scale or contrast Intensity scaling (constant scale or contrast stretched for each cell)stretched for each cell)Feature levels and granularityFeature levels and granularityObjectfeaturesSingleObjectSingleCellSingleFieldCellfeaturesFieldfeaturesGranularity: 2D, 3D, 2Dt, 3DtAggregate/average operatorCell SegmentationCell SegmentationSingle cell segmentation approachesSingle cell segmentation approachesVoronoiVoronoiWatershedWatershedSeeded WatershedSeeded WatershedLevel Set MethodsLevel Set MethodsGraphical ModelsGraphical ModelsVoronoi diagramVoronoi diagramSeedEdgeVertexGiven a set of seeds, draw vertices and edges such that each seed is enclosed in a single polygon where each edge is equidistant from the seeds on either side.Voronoi Segmentation ProcessVoronoi Segmentation Process•Threshold DNA image (downsample?)Threshold DNA image (downsample?)•Find the objects in the imageFind the objects in the image•Find the centers of the objectsFind the centers of the objects•Use as seeds to generate Voronoi diagramUse as seeds to generate Voronoi diagram•Create a mask for each region in the Voronoi Create a mask for each region in the Voronoi diagramdiagram•Remove regions whose object that does not Remove regions whose object that does not have intensity/size/shape of nucleushave intensity/size/shape of nucleusOriginal DNA imageAfter thresholding and removing small objectsAfter triangulationAfter removing edge cells and filteringFinal regions masked onto original imageWatershed SegmentationWatershed SegmentationIntensity of an image ~
View Full Document