二、 文獻探討
4.6 各類型實驗之偵測結果
此部份將對於各項類型影像序列進行偵測，因一個強健的背景模組能以單一門檻來 應付絕大部份的影像序列，且參數統一的優點在於能達到自動化偵測，而不需人工操
作，故以下實驗雖不一定為本方法之最佳解，但皆能以差異不大、甚至完全相同之參數 得到誤差值較低之偵測結果。
表十三 各類型實驗參數表 方法名稱 LFP-SCRG
r 1 R 3
P 4 n 2
TB 0.8 T w 0.001
T s 0.7 T s' 0.7
d 0.2 s 0.8
表十四 偵測結果比較表
編號 實驗名稱 方法 K FN FP FP+FN FPS
LBP 4 5 1595 1600 2.45 LFP 4 0 1080 1080 0.9 1 2 人
LFP-SCRG 1.43 64 462 526 4.83 LBP 4 24 1027 1051 2.6 LFP 4 32 82 114 1.01 2 e 咖水波
LFP-SCRG 1.15 134 69 203 5.09 LBP 4 0 405 405 2.65 LFP 4 14 182 196 1.04 3 e 咖瀑布
LFP-SCRG 1.03 79 40 119 5.1 LBP 4 203 791 994 2.46 LFP 4 136 852 988 0.93 4 階梯
LFP-SCRG 1 470 248 718 5.2
LBP 4 6 1536 1542 2.53 LFP 4 11 1092 1103 0.95 5 綜教跑
LFP-SCRG 1.14 99 580 679 4.65 6 操場 LBP 4 5 1017 1022 2.71
LFP 4 29 526 555 1.2 LFP-SCRG 1.02 43 585 628 5.71 LBP 4 0 182 182 2.69 LFP 4 9 71 80 0.88 7 樹
LFP-SCRG 1 19 51 70 4.79 LBP 4 868 1460 2328 2.64 LFP 4 1447 559 2006 0.97 8 Bootstrap
LFP-SCRG 1.99 1355 1169 2524 4.48 LBP 4 1231 2153 3384 2.62 LFP 4 1262 1900 3162 1.05 9 Camouflage
LFP-SCRG 2.1 5985 514 6499 4.8
LBP 4 1292 924 2216 2.49 LFP 4 1310 1628 2938 0.99 10 ForegroundAperture
LFP-SCRG 1 2363 855 3218 5
LBP 4 578 8093 8671 2.53 LFP 4 1620 2650 3710 0.87 11 LightSwitch
LFP-SCRG 1 845 6031 6876 5.02 LBP 4 0 0 0 2.42 LFP 4 0 0 0 0.84 12 MovedObject
LFP-SCRG 1 0 0 0 5.5 LBP 4 371 263 634 2.33 LFP 4 247 436 683 0.79 13 TimeOfDay
LFP-SCRG 1 674 144 818 4.35 LBP 4 245 1457 1702 0.06 LFP 4 17 1059 1076 0.04 14 車子
LFP-SCRG 1.81 170 972 1142 0.92
編 號
實驗名稱 LBP LFP LFP-SCRG
1 2 人
2 e 咖水波
3 e 咖瀑布
4 階梯
5 綜教跑
6 操場
7 樹
8 Bootstrap
9 Camouflage
10 ForegroundAperture
11 LightSwitch
12 MovedObject
13 TimeOfDay
14 車子
圖三十二 各類型實驗偵測結果圖
五、結論
本研究提出多維度特徵、低運算量、高處理速度、高精確度、動態配置模組數、及 高背景適應力之方法。本方法乃透過多維度特徵提高像素點資訊之完整度，並利用區塊 化及Kernel 函數提升原 LFP 演算法之效能，背景建模方面，透過 SCRG 演算法以動態 配置模組及更新，接著依照模組之權重適時清除重要性低下之模組，最後利用形態學優 化物件輪廓及清除背景雜訊。由實驗結果得知，在單調背景中，本方法展現了最低之模 組數、高精確度之物件面積、運算速度高之優勢；而在複雜背景中，則能有效配置模組 數，令特徵變化大之像素點擁有較多模組，而變化小之像素點仍舊保持最低需求之模組 數，同時能即時訓練規律移動之背景物件至模組中，同時保持一定的運算速度。運算速 度上遠超過LBP 法及原 LFP 法，而本方法於各實驗之平均精確度上，以 FN 值誤差較 高，其導因於形態學之侵蝕作用，而FP 之誤差雖然較低，但亦未能全然改善原紋理特 徵方法之問題，因錯誤點也普遍落在物件輪廓部份，整體而言物件輪廓比其他兩方法較 浮貼於理想遮罩。平均模組數量上勝於兩方法，實驗結果證實了本方法在低平均模組數 下能有極高優勢之偵測精確度及運算速度，達到本研究期望提高基於紋理特徵之移動物 件偵測效能的目標。本方法中含一定比例之色彩特徵，因而在亮度劇烈變化之實驗中，
尚無法完全區別物件之輪廓，然而本方法之背景適應力可以令亮度劇烈變化後，在極短 的影像數(約 3~4 張)內完全將光影訓練至背景中，因此若為即時連續影像，短時間內光 影變化影響被降至最低之效果是可以預期的。
參考文獻
[1] A. M. Elgammal, D. Harwood, and L. S. Davis, “Nonparametric Background Model for Background Subtraction,” in Proceedings of 6th European Conference on Computer Vision, Dublin, Ireland, vol. 2, pp. 751–767, Jun. 2000.
[2] A. Monnet, A. Mittal, N. Paragios, and V. Ramesh, “Background Modeling and Subtraction of Dynamic Scenes,” in Proceedings of 9th IEEE International Conference on Computer Vision, Nice, France, pp. 1305–1312, Oct. 2003.
[3] B. K. P. Horn and B. G. Schunch, “Determining optical flow,” Artificial Intelligence, vol.
17, pp.185-203, 1981.
[4] B.F. Zhang, J. Zhou, and J.C. Zhu, “Research on Three Image Difference Algorithm,” in Proceedings of 2010 International Conference on Image Analysis and Signal Processing, Zhejiang, China, pp. 603-606, Apr. 2010.
[5] C. Zhan, X. Duan, S. Xu, Z. Song, and M. Luo, “An Improved Moving Object Detection Algorithm Based on Frame Difference and Edge Detection,” in Proceedings of 4th International Conference on Image and Graphics, Sichuan, China, pp. 519 – 523, Aug.
2007.
[6] Chen-Sen Ouyang,Chih-Chung Wang,and Yung-Chih Chen, “A Recursive SVD-Based Self-Constructing Rule Generation for Neuro-Fuzzy System Modeling,” 2011 International Conference on Machine Learning and Cybernetics , Guilin, Guangxi, China, 2011.7
[7] D. Culibrk, O. Marques, D. Socek, H. Kalva, and B. Furht, “Neural Network Approach to Background Modeling for Video Object Segmentation,” IEEE Transactions on Neural Networks, vol. 18, no. 6, pp. 1614-1627, 2007.
[8] D. Wang, T. Feng, H. Shum, and S. Ma, “A Novel Probability Model for Background Maintenance and Subtraction,” in Proceedings of 15th International Conference on Vision Interface, Calgary, Canada, pp. 109–116, May 2002.
[9] D.-M. Tsai and S.-C. Lai, “Independent Component Analysis-Based Background Subtraction for Indoor Surveillance,” IEEE Transactions on Signal Processing, vol. 18, no. 1, pp. 158-167, 2009.
[10] D.-Y. Chen, K. Cannons, H.-R. Tyan, S.-W. Shih, and H.-Y.M. Liao, “Spatiotemporal Motion Analysis for the Detection and Classification of Moving Targets,” IEEE Transactions on Multimedia, vol. 10, no. 8, pp. 1578-1591, 2008.
[11] H. Bhaskar, L. Mihaylova, and A. Achim, “Video Foreground Detection Based on Symmetric Alpha-Stable Mixture Models,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 20, no. 8, pp. 1133-1138, 2010.
[12] H. Kim, R. Sakamoto, I. Kitahara, T. Toriyama, and K. Kogure, “Background Subtraction Using Generalised Gaussian Family Model,” Electronics Letters, vol. 44, no.
3, pp. 189-190, 2008.
[13] H. Zhang and M. Zhou, “Objects Tracking Based on Optical Flow in Polar-log Images,”
Journal of Computational Information Systems, pp.829-832, 2006.
[14] J. L. Barron, D. J. Fleet, and S. S. Beauchemin, “Performance of Optical Flow Techniques,” International Journal of Computer Vision, vol. 12, no. 1, pp.43-77, 1994.
[15] J. Schmüdderich, V. Willert, J. Eggert, S. Rebhan, C. Goerick, G. Sagerer, and E.
Körner, “Estimating Object Proper Motion Using Optical Flow, Kinematics, and Depth Information,” IEEE Transactions on Systems, Man, and Cybernetics, Part B, pp.1139-1151, 2008.
[16] J. Stander, R. Mech, and J. Ostermann, “Detection of Moving Cast Shadows for Object Segmentation,” IEEE Transactions on Multimedia, vol. 1, no. 1, pp. 65-76, 1999.
[17] J.M. McHugh, J. Konrad, V. Saligrama, and P.-M. Jodoin,“Foreground-Adaptive Background Subtraction,” IEEE Signal Processing Letters, vol. 16, no. 5, pp. 390-393, 2009.
[18] K. Toyama, J. Krumm, B. Brumitt, and B. Meyers, “Wallflower: Principles and Practice of Background Maintenance,” in Proceedings of 7th IEEE International Conference of Computer Vision, Corfu, Greece, vol. 1, pp. 255–261, Sep. 1999.
[19] L.Y. Liu, N. Sang, and R. Huang, “Background Subtraction Using Shape and Colour Information,” Electronics Letters, vol. 46, no. 1, pp. 41-43, 2010.
[20] Lucia Maddalena and Alfredo Petrosino, “A Self-Organizing Approach to Background Subtraction for Visual Surveillance Applications,” IEEE Transactions on Image Processing, vol. 17, no. 7, July 2008.
[21] M. Heikkilä and M. Pietikäinen,“A Texture-based Method for Modeling the Background and Detecting Moving Objects,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 28, no. 4, pp. 657–662, 2006.
[22] M. Surucu, D. Surucu, and R. Hacioglu, “Detecting and Tracking Moving Objects in Real Time Images via Active Camera,” in Proceedings of IEEE 18th Signal Processing and Communications Applications Conference, Diyarbakir, Turkey, pp. 764-767, Apr.
2010.
[23] M. Yeasin, “Optical Flow in Log-mapped Image Plane-A New Approach,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, no.1, pp.125-131, 2002.
[24] P.-M. Jodoin, M. Mignotte, J. Konrad, “Statistical Background Subtraction Using Spatial Cues,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 17, no. 12, pp. 1758-1763, 2007.
[25] R. Cucchiara, C. Grana, M. Piccardi, and A. Prati, “Detecting Moving Objects, Ghosts, and Shadows in Video Streams,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 10, pp. 1337–1342, 2003.
[26] R. Cucchiara, M. Piccardi, and A. Prati, “Detecting moving objects,ghosts, and shadows in video streams,” IEEE Trans. Pattern Anal. Mach. Intell, vol. 25, no. 10, pp. 1–6, Oct.
2003.
[27] R. T. Collins, A. J. Lipton, and T. Kanade, “A System for Video Surveillance and Monitoring,” Technical Report, CMU-RI-TR-00-12, Robotics Institute, Carnegie Mellon University, May 2000.
[28] R. Zhang, S. Zhang, and S. Yu, “Moving Objects Detection Method Based on Brightness Distortion and Chromaticity Distortion,” IEEE Transactions on Consumer Electronics, vol. 53, no. 3, pp. 1177-1185, 2007.
[29] S. Negahdaripour, “Revised Definition of Optical Flow: Integraion of Radiometric and Geometric Cues for Dynamic Scene Analysis,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 20, no. 9, pp. 961–979, 1998.
[30] S. Wei, Z. Chen, M. Li, L. Zhuo, “An Improved Method of Motion Detection Based on Temporal Difference,” in Proceedings of 2009 International Workshop on Intelligent Systems and Applications, Wuhan, China, pp. 1-4, May 2009.
[31] S.-S. Huang, L.-C. Fu, and P.-Y. Hsiao, “Region-Level Motion-Based Background Modeling and Subtraction Using MRFs,” IEEE Transactions on Image Processing, vol.
16, no. 5, pp. 1446-1456, 2007.
[32] W. Wang, J. Yang, and W. Gao, “Modeling Background and Segmenting Moving Objects from Compressed Video,” IEEE Transactions on Circuit and Systems for Video Technology, vol. 18, no. 5, pp. 670-681, 2008.
[33] Y.-J. Du and Y.-J. Jin, “Log-polar Coordinate Domain Gradient Optical Flow Tracking
[34] 劉佳諺, “具強健性之基於紋理的背景相減法於影像中移動物件偵測,” 義守大學資訊工 程系碩士論文, 2007。