Beispielpräsenation

Computer aided navigation on histological whole slide images: 

N. Zerbe, M. Flügge, P. Hufnagl , K. Schlüns InstitutE of Pathology , Charité – Universitätsmedizin Berlin Computer aided navigation on histological whole slide images

Introduction: 

Introduction Motivation Context based navigation on virtual slides Laborious and time- consuming tracking of „ diagnostic clues “ over serial sections Functionality enhancement of conventional virtual microscopes Goal Computer assisted navigation based on automated image analysis of whole slide images (WSI)

Material & Methods (1): 

Material & Methods (1) Degrees of complexity Material Preparation artifacts Biopsy versus surgical specimen Methods Tile versus WSI based processing Intra versus inter slide tracking Intra versus inter stain tracking Focus Biopsies WSI Intra slide tracking Rigid tissue model

Material & Methods (2): 

Material & Methods (2) Preprocessing Segmentation of tissue particles Reconstruction of serialization curve Identification of serial sections Classification Rotation and translation invariant contour based features Expectation maximization classifier Reconstruction of particle sequence

Material & Methods (3): 

Material & Methods (3) Alignment ROI based extraction and mapping of key points Calculation of transformations for navigation Offline/Online Processing case , slide , serial sections , particles Key point pairs Transformations

Results (1): 

Results (1)

Results (2): 

Results (2) http://virtuellemikroskopie.charite.de/dgp2010nav/

Results (3): 

Results (3)

Future Work: 

Future Work Tracking across multiple slides and stains Enhancement related to surgical specimens Provide algorithms via web services

Conclusion: 

Conclusion Tracking of „ diagnostic clues “ across serial sections by retaining a relative position Time efficient navigation in serial sections of histological slides New diagnostic utility which does not exist in conventional virtual microscopes

PowerPoint Presentation: 

Thank you for your attention ! This work has been supported by Zukunftsfonds Berlin (TSB, IBB).

Animationstest: 

Animationstest Text 1 Text 3 Text 2