臺灣博碩士論文加值系統

物件追蹤是一門重要的研究主題，現今許多應用廣泛使用到物件追蹤技術如：車輛追蹤、人臉追蹤或運動估測…等。在電腦視覺中所提出物件追縱的方法非常多，而本論文所使用的光流法(optical flow method)應用於在物件追蹤上，雖其過程需要較大的計算量，但相對地，物件追蹤的正確性與環境適應性較能得到顯見的效果。本論文研究方向主要目的是以Lucas-Kanade光流法作為物件追蹤的基礎，Lucas-Kanade光流法主要求出一組適當的參數用於追蹤物件上，而本論文藉由使用粒子群聚最佳化(particle swarm optimization, PSO)針對Lucas-Kanade光流法中的參數做最佳化處理，藉由PSO所估計出適當的參數範圍，能有效的限制粒子群飛行的範圍，適時地取代傳統的區域求解的作法，預期能更有效地達到物件追蹤的目的。

The problem of object tracking is an important research topic. Object tracking is used in many applications such as vehicle tracking, face tracking, and motion estimation. In computer vision, there are many methods for object tracking. This research uses optical flow method for object tracking. Although this method is time consuming, it is effective in accuracy of tracking and environment adaptive. This study focuses on Lucas-Kanade optical flow. In order to improve the executing time, particle swarm optimization is used to compute a suitable range of parameters and reduce the search range. The purposed method outperforms the traditional method for local solutions, and can effectively achieve the goal of object tracking.

目錄
圖目錄II
摘要IV
ABSTRACTV
第1章 研究動機與文獻探討1
第2章 Lucas-Kanade光流法3
2.1 區塊匹配方式及內插法4
2.2 最陡梯度影像及Hessian矩陣推導6
2.3 參數更新8
第3章 Lucas-Kanade光流法實驗觀察9
3.1 實驗方法9
3.2 實驗數據9
第4章 粒子群聚最佳化14
第5章 PSO-based LK optical flow16
5.1 PSO與Lucas-Kanade光流法之結合構想16
5.2 PSO-based LK架構方式16
第6章 PSO-based LK實驗結果19
6.1 實驗作法19
6.2 旋轉(rotation)20
6.3 位移(translation)22
6.4 縮放(zoom)24
6.6 車輛追蹤28
第7章 結論32
參考文獻33
圖目錄
圖2.1、Lucas-Kanade光流法應用於物件追蹤3
圖2.2、當t=0從I0取出的樣版影像T4
圖2.3、當t=n時必須先轉回與原本T的相同大小，Ṫ才能作比較4
圖2.4、線性內插法表示圖5
圖2.5、(a)圖為垂直梯度影像；(b)圖為水平梯度影像6
圖2.6、最陡梯度影像(steepest descent image)實驗結果圖7
圖2.7、Lucas-Kanade光流法之實作流程圖8
圖3.1、追蹤物件平移情況實驗9
圖3.2、實驗一之結果發現參數p中p5和p6變化遠大於其餘各參數10
圖3.3、實驗一中參數p的p1~p4各參數變化情形10
圖3.4、追蹤物件旋轉情況實驗11
圖3.5、追蹤物件旋轉情況下參數p中p5和p6依然比其他參數還大11
圖3.6、追蹤物件旋轉情況下，p1~p4所變化非常規律12
圖3.7、追蹤物件縮放情況實驗12
圖3.8、追蹤物件縮放時參數p5及p6的變化情形13
圖3.9、追蹤物件縮放時參數p1~p4的變化情形13
圖4.1、粒子群聚最佳化演算法之流程圖15
圖5.1、使用PSO避免Lucas-Kanade光流法中參數p出現發散情形16
圖5.2、PSO使用於Lucas-Kanade光流法之流程圖18
圖6.1、當追蹤物件旋轉時PSO-based LK之實驗結果圖20
圖6.2、當追蹤物件旋轉時PSO-based LK與Lucas-Kanade光流法MSE值比較21
圖6.3、當追蹤物件位移時PSO-based LK之實驗結果圖22
圖6.4、當追蹤物件位移時PSO-based LK與Lucas-Kanade光流法之MSE值23
圖6.5、當追蹤物件位移時PSO-based LK與Lucas-Kanade光流法的參數p5變化23
圖6.6、當追蹤物件位移時PSO-based LK與Lucas-Kanade光流法的參數p6變化23
圖6.7、當追蹤物件縮放時，PSO-based LK之實驗結果圖24
圖6.8、當追蹤物件縮放時PSO-based LK與Lucas-Kanade光流法之MSE值比較25
圖6.9、當追蹤物件縮放時PSO-based LK與Lucas-Kanade光流法之參數p5比較25
圖6.10當追蹤物件縮放時PSO-based LK與Lucas-Kanade光流法之參數p6比較25
圖6.11、當物件被遮蔽時Lucas-Kanade光流法之實驗結果圖26
圖6.12、當物件被遮蔽時，PSO-based LK之實驗結果圖27
圖6.13、當物件被遮蔽時PSO-based LK與Lucas-Kanade光流法之MSE值比較27
圖6.14、在交通車輛影片上，Lucas-Kanade光流法實驗結果28
圖6.15、在交通車輛影片上PSO-based LK之實驗結果圖29
圖6.16、交通車輛影片中PSO-based LK與Lucas-Kanade光流法之MSE值比較29
圖6.17、在交通車輛影片上Lucas-Kanade光流法之參數p1~p4變化30
圖6.18、在交通車輛影片上Lucas-Kanade光流法之參數p5與p6變化30
圖6.19、在交通車輛影片上，PSO-based LK之參數p1~p4變化31
圖6.20、在交通車輛影片上，PSO-based LK之參數p5與p6變化31

[1]A. Sánchez, J. V. Ruiz, A. B. Moreno, A. S. Montemayor, J. Hernández, and J.J. Pantrigo, “Differential optical flow applied to automatic facial expression recognition,” in Press Corrected Proof, Available online 20, Oct., 2010.
[2]B. D. Lucas and T. Kanade, “An iterative image registration technique with an application to stereo vision,” application to stereo vision, in: Proc. of the International Joint Conference on Artificial Intelligence, Canada, pp. 674–679, Aug., 1981.
[3]Berthold K.P. Horn and Brian G. Schunck, “Determining optical flow,” Artifcial Intelligence, pp. 185-203, 1981.
[4]D. Cobzas, M. Jagersand, and P. Sturm, “3D SSD tracking with estimated 3D planes,” image and vision computing , pp. 69-79, 2009.
[5]D. J. Fleet and A. D. Jepson, “Computation of component image velocity from local phase information,” International Journal of Computer Vision, vol. 5, no. 1, pp. 77-104, 1990.
[6]D Schreiber, “Generalizing the Lucas-Kanade algorithm for histogram-based tracking,” Pattern Recognition Letters, vol. 29, vol. 7, pp. 852-861, May., 2008.
[7]D Schreiber, “Incorporating symmetry into the Lucas-Kanade framework,” Pattern Recognition Letters, vol. 30, no. 7, pp. 690-698, May., 2009.
[8]J. Kennedy, and R. C. Eberhart, “Particle swarm optimization,” in: Proc. IEEE Int. Conf. on Neural Networks, Perth, Australia, vol. 4, pp. 1942-1948, 1995.
[9]J. Zhu, M. R. Lyu, and T. S. Huang, “A fast 2D shape recovery approach by fusing features and appearance,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, no. 7, pp. 1210-1224, Jul., 2009.
[10]Heeger D. J, “Optical flow using spatiotemporal filters,” International Journal of Computer Vision, vol. 1, no. 4, pp. 279-302, 1987.
[11]H. H. Nagel, “On the estimation of optical flow: relations between different approaches and some new results,” Artifcial Intelligence, vol. 33, no. 3, Nov., 1987.
[12]L. Matthews, T. Ishikawa, and S. Baker, “The template update problem,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 26, no. 6, pp. 810-815, Jun., 2004.
[13]N. Dowson and R. Bowden, “Mutual information for Lucas-Kanade tracking (MILK): an inverse compositional formulation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 30, no. 1, pp. 180-185, Jan., 2008.
[14]P. Anandan, “A computational framework and an algorithm for the measurement of visual motion,” International Journal of Computer Vision, vol. 2, no. 3, pp. 283-310, 1988.
[15]P. Sungsoo and K. Daijin, “Subtle facial expression recognition using motion magnification,” Pattern Recognition Letters, vol.30, no. 7, pp. 708-716, May., 2008.
[16]R. C. Eberhart, and J. Kennedy, “A new optimizer using particle swarm theory,” in: Proc. IEEE Int. Symposium on Micro Machine and Human Science, Nagoya, Japan, pp. 39-43, 1995.
[17]S. Baker and I. Mattews, “Lucas-Kanade 20 years on: a unifying framework,” International Journal of Computer Vision, vol. 56, no. 3, pp. 221-255, 2004.
[18]S. Lim, J. G. Apostolopoulos, and A. E. Gamal, “Optical flow estimation using temporally oversampled video,” IEEE Transactions on Image Processing, vol. 14, no. 8, pp. 1074-1087, Aug., 2005.
[19]T. Marco, “A genetic algorithm for optical flow estimation,” Image and Vision Computing, vol. 25, no. 2, pp. 141-147, Jan., 2006.