臺灣博碩士論文加值系統

視訊物件追蹤在電腦視覺的研究領域中，扮演著相當重要的角色，例如視訊監視系統、知覺使用者介面或是物件式影片壓縮等。視訊物件追蹤的困難點有很多，像是複雜的背景、被遮蔽的物體、場景的光影變化、或是會變形的物體。這些情況都會讓視訊物件追蹤的工作變得困難許多。
在本篇論文中，我們假設在特徵點周圍的色彩分佈是重要的，並提出一個基於這個想法的強韌色彩統計分佈。由實驗結果可以看出，在某些困難的情況下，像是場景的光影變化或是被遮蔽的物體，我們所提出的強韌色彩統計分佈表現的很不錯。
此外，我們利用這個強韌色彩統計分佈來當作物件的外觀模型，以協助我們去做好視訊物件追蹤的工作。我們修改傳統的粒子濾波架構，並且把每一個偵測到的特徵點當作是粒子，進而發展出一個新的物件追蹤演算法。由實驗結果可以看出，我們所提出的物件追蹤演算法相當令人滿意。

Video object tracking is a very important issue in computer vision applications, such as video surveillance, perceptual user interfaces, and object-based video compression. The difficulty of video object tracking might come from many factors, such as cluttered background, occlusions, lighting changes and deformation.
In this thesis, we assume that color features in the neighborhood of interest points are important, and propose to use a robust color histogram based on SURF. As shown in the experimental results, the robust color histogram performs well in several difficult scenarios, such as lighting changes and occlusions.
Moreover, we use the robust color histogram as the appearance model of the target object to assist video object tracking. We modify the traditional particle filter framework and view each SURF interest point as a particle to develop a new object tracking algorithm. In the experimental results, we have shown that the performance of our object tracking algorithm is good.

1 Introduction
1.1 Difficulties
1.2 Related Work
1.2.1 Region-based methods
1.2.1.1 Level-set methods
1.2.1.2 Segmentation methods
1.2.2 Object-based methods
1.2.2.1 Deterministic methods
1.2.2.2 Stochastic methods
1.2.2.3 Hybrid deterministic and stochastic methods
2 Object Representations
2.1 SURF: Speeded Up Robust Features
2.2 Robust Color Histograms
2.3 Experimental Results
3 Object Tracking Using Particle Filters
3.1 Background of Particle Filters
3.1.1 States and Observations
3.1.2 The State Space Model
3.1.3 Problem Formulation
3.1.4 Monte Carlo Simulation
3.1.5 Importance Sampling
3.1.6 State Estimates
3.1.7 General Assumptions
3.1.8 The Degeneracy Problem
3.1.9 The Particle Filtering Algorithm
3.2 Background of Auxiliary Particle Filters
3.2.1 Motivation
3.2.2 The Auxiliary Particle Filtering Algorithm
3.3 Proposed Methods
3.3.1 SURF Particle Filter
3.3.1.1 Motivation
3.3.1.2 Implementation Issues
3.3.1.3 SURF Particle Filter Algorithm
3.3.2 SURF Particle Filter Algorithm using Robust Color Histogram
3.3.2.1 Motivation
3.3.2.2 Implementation Issues
3.3.2.3 SURF Particle Filter using Robust Color Histograms Algorithm
4 Experimental Results
5 Conclusions
5.1 Contribution
5.2 Future work
6 REFERENCES

[1] A. R. Mansouri, “Region Tracking via Level Set PDEs without Motion Computation,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 7, pp. 947-961, July 2002.
[2] A. Yilmaz, X. Li, and M. Shah, “Contour-based Object Tracking with Occlusion Handling in Video Acquired Using Mobile Cameras,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 26, no. 11, pp. 1531-1536, November 2004.
[3] V. Mezaris, I. Kompatsiaris, and M. G. Strintzis, “Video Object Segmentation Using Bayes-Based Temporal Tracking and Trajectory-Based Region Merging,” IEEE Trans. Circuits and Systems for Video Technology, vol. 14, no. 6, pp. 782-795, June 2004.
[4] P. Withagen, K. Schutte, and F. Groen, “Probabilistic Classification Between Foreground Objects and Background,” Proc. 17th Int’l Conf. Pattern Recognition, vol. 1, pp. 31-34, Aug 2004.
[5] D. Comaniciu, V. Ramesh, and P. Meer, “Kernel-Based Object Tracking,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 25, no. 5, pp. 564-577, May 2003.
[6] D. Comaniciu, and P. Meer, “Mean Shift: A Robust Approach toward Feature Space Analysis,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 24, no. 5, pp. 603-619, May 2002.
[7] T. L. Liu, and H. T. Chen, “Real-Time Tracking Using Trust-Region Methods,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 26, no. 3, pp. 397-402, March 2004.

[8] R. E. Kalman, “A New Approach to Linear Filtering and Prediction Problems,” Transactions of the ASME - Journal of Basic Engineering, vol. 82, pp. 35-45, March 1960.
[9] K. Nummiaro, E. Koller-Meier, and L. V. Gool, “An Adaptive Color-Based Particle Filter,” Image and Vision Computing, vol. 21, pp. 99-110, 2003.
[10] K. Deguchi, O. Kawanaka, and T. Okatani, “Object tracking by the mean-shift of regional color distribution combined with the particle-filter algorithm,” Proc. 17th Int’l Conf. Pattern Recognition, vol. 3, pp. 506-509, Aug 2004.
[11] E. Maggio and A. Cavallaro, “Hybrid particle filter and mean shift tracker with adaptive transition model,” Proc. Int’l Conf. Acoustics, Speech, and Signal Processing, March 2005.
[12] H. Bay, T. Tuytelaars, and L. V. Gool, “SURF: Speeded Up Robust Features,” Proc. 9th European Conf. on Computer Vision, May 2006.
[13] P. Viola and M. Jones, “Rapid object detection using a boosted cascade of simple feature,” Proc. IEEE Conf. on Computer Vision and Pattern Recognition, 2001.