1. Learning to Compute the Symmetry Plane for Human Faces Jia Wu (jiawu@uw.edu) ACM-BCB '11, August 2011 1

2. A simple introduction for classification 2 Positive /negative samples Classifier testing Training

3. Landmark by medical experts 3 Landmarks labeled by experts Standard symmetry plane

4. Flow chart for training 4 Landmarks from medical experts Landmark model training Points predicted as interesting regions Connected regions Standard symmetry plane Symmetry model training

5. 10 kinds of landmarks. – Nose: ac (nose side), prn, sn,se – Eyes: en, ex – Mouth: (li,ls), ch, sto – Chin: slab 5

6. Positive/negative samples 6 Training for en: the inner corners of the eyes Training for prn: most protruded point of nasal tip

7. Features: mean and Gaussian curvatures for original head and smoothed head 7

8. Flow chart 8 Landmarks from medical experts Landmark model training Points predicted as interesting Connected regions Standard symmetry plane Symmetry model training

9. Interesting points prediction 9 Prediction of en: the inner corners of the eyes Prediction of prn: most protruded point of nasal tip

10. Connected regions 10 Connected regions for en: each color means one region Connected regions for prn: each color means one region

11. Flow chart 11 Landmarks from medical experts Landmark model training Points predicted as interesting regions Connected regions Standard symmetry plane Symmetry model training

12. How to define “good” symmetric regions A “good” pair of regions should be symmetric to the standard symmetry plane A “good” single region should have the center on the standard symmetry plane 12 “good” regions for en “good” regions for prn

13. Feature for regions 13 Principal component analysis and eigenvalues

14. Flow chart for training 14 Landmarks from medical experts Landmark model training Points predicted as interesting regions Connected regions Standard symmetry plane Symmetry model training

15. Procedure for New data 15 Select possible landmark areas( from Landmark model) Find and pair connected regions Determine good singles and good pairs ( from Symmetry model) Get center and draw a plane using learned centers interesting regions for prn Predicted as good single Predicted as good pair

16. Procedure for New Images 16 Centers of good regionsCenters for constructing plane of symmetry Result: Plane of symmetry

17. Experiments Compare the plane of symmetry to – Ground truth (plane determined by expert labeled landmarks) – Mirror method in literature Ground truth dataset 1 Ground truth dataset 2 Cleft dataset 17

18. Mirror method in literature 18

19. Results compare to ground truth 19 Yellow: overlapping with ground truth Green: ground truth Purple: extra from each method our method Angle:4.03˚ mirror method Angle:2.15˚

20. Results compare to ground truth 20 Yellow: overlapping with ground truth -> true positive Green: ground truth -> false negative Purple: extra from each method -> false positive our methodmirror method

21. 21 best worst best

22. Ground truth dataset 2 12345 678910 22

23. 0306090 0 1247 30 35811 60 691214 90 10131516 23

24. 24 best worst best

25. 25

26. Cleft dataset 26 Learning method Mirror method

27. Questions? 27