1. COMPUTER VISION Tam Lam

2. Tracking: General idea
Initialize model in the first frame
Given model estimate for frame t-1:
Predict for frame t
Use dynamics model of how the model changes
Correct for frame t
Use observations from the image
predict
correct

3. Motion Tracking Deciding on the structure of the model points curves
pictorial structures

4. Tracking issues Deciding on the structure of the model Initialization
Specifying the dynamics model
Specifying the observation model
Data association problem: which measurements tell us about the object(s) being tracked?

5. Data association
Simple strategy: only pay attention to the measurement that is “closest” to the prediction

6. Data association
Simple strategy: only pay attention to the measurement that is “closest” to the prediction
Doesn’t always work…

7. Data association
Simple strategy: only pay attention to the measurement that is “closest” to the prediction
More sophisticated strategy: keep track of multiple state/observation hypotheses
This is a general problem in computer vision, there is no easy solution

8. Tracking issues Deciding on the structure of the model Initialization
Specifying the dynamics model
Specifying the observation model
Data association problem
Prediction vs. correction
If the dynamics model is too strong, will end up ignoring the data
If the observation model is too strong, tracking is reduced to repeated detection
Drift

9. DRIFT
the model is trying to predict, change over time in unforeseen ways
the predictions become less accurate as time passes  problem

10. Drift

11. Tracking with person-specific appearance models
Tracker
pictorial structure

12. Tracking with person-specific appearance models
Structure and dynamics are generic, appearance is person-specific
Trying to acquire an appearance model “on the fly” can lead to drift
Instead, can use the whole sequence to initialize the appearance model and then keep it fixed while tracking
Given strong structure and appearance models, tracking can essentially be done by repeated detection (with some smoothing)

13. Bottom-up initialization: Clustering

14. Top-down initialization: Exploit “easy” poses

15. Tracking by model detection

16. QUESTIONS????