1. AAM based Face Tracking with Temporal Matching and Face Segmentation Mingcai Zhou 1 、 Lin Liang 2 、 Jian Sun 2 、 Yangsheng Wang 1 1 Institute of Automation Chinese Academy of Sciences, Beijing, China 2 Microsoft Research Asia Beijing, China CVPR 2010

2. 2 Outline AAM Introduction Related Work Method and Theory Experiment

3. 3 AAM Introduction A statistical model of shape and grey- level appearance Shape model Appearance model

4. 4 Shape Model Building :mean shape :shape bases,shape parameters learn by PCA generate mean shape 、 shape bases

5. 5 Texture Model Building :mean appearance :appearance bases :appearance parameters W(x) 灰階值 Shape-free patch Mean shape

6. 6 AAM Model Building

7. 7 AAM Model Search Find the optimal shape parameters and appearance parameters to minimize the difference between the warped-back appearance and synthesized appearance map every pixel x in the model coordinate to its corresponding image point

8. 8 Problems- AAM tracker Difficultly generalize to unseen images Clutterd backgrounds

9. 9 How to do? A temporal matching constraint in AAM fitting -Enforce an inter-frame local appearance constraint between frames Introduce color-based face segmentation as a soft constraint

10. 10 Related Work - feature-based (mismatched local feature) Integrating multiple visual cues for robust real-time 3d face tracking, W.-K. Liao, D. Fidaleo, and G. G. Medioni. 2007 - intensity-based (fast illumination changes) Improved face model fitting on video sequences, X. Liu, F. Wheeler, and P. Tu. 2007 temporal matching constraint

11. 11 Method and Theory Extend basic AAM to Multi-band AAM –The texture(appearance) is a concatenation of three texture band values The intensity (b) X-direction gradient strength (c) Y-direction gradient strength (d)

12. 12 1.Select feature points with salient local appearances at previous frame 2.I ( t−1) to the Model coordinate and get the appearance A (t-1) 3.Use warping function W(x;p t ) maps R (t-1) to a patch R (t) at frame t Temporal Matching Constraint

13. 13 Shape parameter Initialization, Face Motion Direction

14. 14 Shape parameter Initialization When r reaches the noise level expected in the correspondences, the algorithm stops

15. 15 Shape parameter Initialization -Comparison Motion direction Feature matching Previous frame’s shape

16. 16 Face Segmentation Constraint Where are the locations of the selected outline points in the model coordinate

17. 17 Face Segmentation Constraint -Face Segmentation

18. 18 Face Segmentation Constraint

19. 19 Experiments Lost frame num

20. 20 Experiments

21. 21 Conclusion ─ Our tracking algorithm accurately localizes the facial components, such as eyes, brows, noses and mouths, under illumination changes as well as large expression and pose variations. ─ Our tracking algorithm runs in real-time. On a Pentium-4 3.0G computer, the algorithm’s speed is about 50 fps for the video with 320 × 240 resolution

22. 22 Future Work ─ Our tracker cannot robustly track profile views with large angles ─ The tracker’s ability to handle large occlusion also needs to be improved