1. Kaihua Zhang Lei Zhang (PolyU, Hong Kong) Ming-Hsuan Yang (UC Merced, California, U.S.A. ) Real-Time Compressive Tracking

2. Introduction Random projection Classifier construction and update Experiments Conclusion 2

3. Introduction Random projection Classifier construction and update Experiments Conclusion 3

4. Propose an effective and efficient tracking algorithm with an appearance model based on features extracted in the compressed domain. Our appearance model employs non-adaptive random projections that preserve the structure of the image feature space of objects Compress samples of foreground targets and the background using the same sparse measurement matrix The tracking task is formulated as a binary classification via a naive Bayes classifier with online update in the compressed domain 4

5. 5 Random Measurement Matrix high-dimensional space lower-dimensional space

6. 6

7. 7

8. Introduction Random projection Classifier construction and update Experiments Conclusion 8

9. 9

10. 10

11. 11 where d is the original and k the reduced dimensionality of the data set

12. 12

13. 13 Selecting matrix A that provide the desired result are:

14. 14 Construct random matrix R such that JL lemma

15. In this work, we set s = m/4 which makes a very sparse random matrix For each row of R, only about c, c ≤ 4, entries need to be computed Therefore, the computational complexity is only O(cn) which is very low Furthermore, we only need to store the nonzero entries of R which makes the memory requirement also very light. 15

16. 16

17. Haar-like features are digital image features used in object recognition A simple rectangular Haar-like feature can be defined as the difference of the sum of pixels of areas inside the rectangle 17

18. 18 Negative entry Positive entry Zero entry

19. 19

20. Introduction Random projection Classifier construction and update Experiments Conclusion 20

21. 21

22. 22

23. 23

24. Introduction Random projection Classifier construction and update Experiments Conclusion 24

25. 25 http://www4.comp.polyu.ed u.hk/~cslzhang/CT/CT.htm

26. 26

27. 27

28. Introduction Random projection Classifier construction and update Experiments Conclusion 28

29. In this paper, we proposed a simple yet robust tracking algorithm with an ap- pearance model based on non-adaptive random projections that preserve the structure of original image space Numerous experiments with state-of-the-art algorithms on challenging sequences demonstrated that the proposed algorithm performs well in terms of accuracy, robustness, and speed. 29