1 / 41

Segmentation of the left atrial appendage from 3D images

Segmentation of the left atrial appendage from 3D images. Pol Grasland-Mongrain 20/04/2009 – 28/08/2009. Views of the Left Atrial Appendage. Views of the Left Atrial Appendage. Variable shapes 1 to 19 cm 3 Function ? Has to be ablated sometimes. Motivation. Current implementation:

faraji
Download Presentation

Segmentation of the left atrial appendage from 3D images

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Segmentation of the left atrial appendage from 3D images PolGrasland-Mongrain 20/04/2009 – 28/08/2009

  2. Views of the Left Atrial Appendage

  3. Views of the Left Atrial Appendage • Variable shapes • 1 to 19 cm3 • Function ? • Has to be ablated sometimes

  4. Motivation • Current implementation: • LAA is represented as a short trunk in the model • Current framework not flexible enoughto grow into highly variable shape

  5. Motivation • Motivation of my master thesis: • Addition of an Automatic Segmentation • Algorithm of the Left Atrial Appendage • in the Philips Framework

  6. Plan • Current Method at Philips • Actual Work • Results and Future Work

  7. 1. HeartDetection 2. Parametric Adaptation(Similarity) 3. Parametric Adaptation(Piecewise Affine) 4. Deformable Adaptation Philips Aachen method New Image Segmentation Chain Segmented Image

  8. 1. HeartDetection 2. Parametric Adaptation(Similarity) 3. Parametric Adaptation(Piecewise Affine) 4. Deformable Adaptation Philips Aachen method New Image Segmentation Chain Segmented Image vadap = T[v] E = Eext[T]

  9. Parametric Adaptation,Deformable Models • Use External Energy :

  10. 1. HeartDetection 2. Parametric Adaptation(Similarity) 3. Parametric Adaptation(Piecewise Affine) 4. Deformable Adaptation Philips Aachen method New Image Segmentation Chain Segmented Image vadap = T[v] E = Eext[T] Free motion for vadap E = Eext[v]+ αEint[v]

  11. Deformable Models • Internal Energy

  12. Plan • Current Method at Philips • Actual Work • Segment manually 17 patients LAA • Modify Philips models • Interface Left Atrium - Left Atrium Appendage • Mesh which inflate • Code an automatic mesh-inflation algorithm • External Energy • Threshold between LAA – Background • Internal Energy • Results and Future Work

  13. Plan • Current Method at Philips • Actual Work • Segment manually 17 patients LAA • Modify Philips models • Interface Left Atrium - Left Atrium Appendage • Mesh which inflate • Code an automatic mesh-inflation algorithm • External Energy • Threshold between LAA – Background • Internal Energy • Results and Future Work

  14. Plan • Current Method at Philips • Actual Work • Segment manually 17 patients LAA • Modify Philips model • Interface Left Atrium - Left Atrium Appendage • Mesh which inflate • Code an automatic mesh-inflation algorithm • External Energy • Threshold between LAA – Background • Internal Energy • Results and Future Work

  15. Model Modification

  16. Plan • Current Method at Philips • Actual Work • Segment manually 17 patients LAA • Modify Philips model • Interface Left Atrium - Left Atrium Appendage • Mesh which inflate • Code an automatic mesh-inflation algorithm • External Energy • Threshold between LAA – Background • Internal Energy • Results and Future Work

  17. External and Internal Energies Edge- based Region-based Mesh Reference Triangle Regularization Curvature Internal Energy External Energy N-Gon Regularization

  18. External Energy : Edge-Based • No specific features

  19. External Energy : Region-Based • Gray Value Above or Under ?

  20. External Energy : Region-Growing • Gray Value Still Above (Under) ? • Already Annotated ?

  21. External Energy : Region-Growing • Gray Value Still Above (Under) ? • Already Annotated ? Gray Value Still Above (Under) ? Already Annotated ?

  22. External Energy : Region-Growing • Gray Value Still Above (Under) ? • Already Annotated ? Gray Value Still Above (Under) ? Already Annotated ?

  23. Threshold LAA-Myocardium

  24. Threshold LAA-Myocardium

  25. Threshold LAA-Myocardium • Minimization of classification error • Stop when Area1 = Area2

  26. Internal Energy : Mesh Reference • Updated Mesh

  27. Internal Energy : Triangle Regularization • Approximate each triangle by a rotated and scaled equilateral triangle

  28. Internal Energy : Curvature • Remove the peaks

  29. Internal Energy : Curvature • Remove the peaks

  30. Internal Energy : N-Gon Regularization • Approximate each “N-Gon” by a rotated and scaled regular N-Gon

  31. External and Internal Energies Edge- based Region-based Mesh Reference Triangle Regularization Curvature Internal Energy External Energy N-Gon Regularization

  32. Plan • Current Method at Philips • Actual Work • Results and Future Work

  33. Results

  34. Results • Main problem : loops -> repair

  35. Results Specificity = True Pos. / (True Pos. + False Neg.) Quality = True Pos. / (True Pos. + False Pos.)

  36. Results • Sum up: • almost all segmented voxels really belong to LAA • but the mesh doesn’t inflate enough

  37. Results (1) (10) (7) (5)

  38. Majors Fails (11) (14)

  39. Possible future works • Improve the loop repair : • Freeze vertices • Better correction • Find a new internal energy ?

  40. Thank you for your attention ! Any Questions ?

More Related