本論文提出一個基於輪廓模糊類神經網路之移動物體辨識系統。在擷取移動物體方面，我們用了一連串的影像處理方法;這些方法包括:現有影像與背景影像之灰階值相減，Sobel運算，型態學處理。在獲取物體的特徵值方面，我們運用輪廓為基礎的模型。使用了輪廓追蹤以及離散傅立葉轉換(DFT)來獲取輪廓的特徵值。另外一個特徵值是長寬比，這個數值可以從擷取物體的垂直投影和水平投影獲得。最後我們使用自我架構類神經模糊推論網路(SONFIN)來訓練以及辨識移動物體。實驗結果顯示我們可以精確的分辨出行人、機車、汽車及狗四種移動物體，同時使用SONFIN的結果比使用類神經網路的辨識結果來得好。

Moving object recognition by contour-based neural fuzzy network is proposed in this thesis. To extract a moving object, we use a series of image processes, including gray-based subtraction between current and background images, Sobel operation, and morphological operation. To extract object’s feature vector, we use contour-based model. Parts of the features are obtained by contour following followed by Discrete Fourier Transform (DFT). Another feature is length-width ratio, which can be derived from vertical and horizontal projection of the extracted object. Finally, we use the Self-Constructing Neural Fuzzy Inference Network (SONFIN) to train and recognize moving objects. The experiment shows we can recognize four moving objects, including a pedestrian, a motorcycle, a car, and a dog, exactly. The performance of SONFIN is shown to be better than a neural network from comparison.

Contents
Chinese Abstract ……………………………………………………….……i
English Abstract…………………………………………………………….ii
Acknowledgments…………………………………………………....….…iii
Contents…….…..…………………………………………………iv
List of Figures…………………………………………………………….....vi
List of Tables……...………………………………………………………viii
Chapter 1 INTRODUCTION………………..………………………1
1.1 Survey………………………………….……………….………1
1.2 Literature Review …………………………..………..………2
1.3 Organization of The Thesis…………………………..………..4
Chapter 2 MOVING OBJECT EXTRACTION…..……….………5
2.1 Overview……………….……………………….…………5
2.2 Image Model………..…..………………………………5
2.3 Gray-Based Subtraction……………………………………….6
2.4 Edge-Based Subtraction………………………………..…….8
2.5 Morphological Operators………………………………………..….11
2.6 Concluding Remarks..……………………………………………...14
Chapter 3 FEATURES EXTRACTION…………………15
3.1 Overview……….………..………….……….…………15
3.2 Length-Width Ratio Calculation…..………………………16
3.3 Contour Feature Extraction………………………..………………19
3.4 Concluding Remarks…………………………………………..26
Chapter 4 TRAINING AND RECOGNITION BY SONFIN……....27
4.1 Overview…………….…………..……………….….……27
4.2 Architecture of SONFIN…………………………….….28
4.3 Structure and Parameter Learning of SONFIN………………33
4.4 Training and Recognition by SONFIN…………………….35
4.5 Concluding Remarks….…………………………………………..36
Chapter 5 SIMULATION………………………………..…….38
5.1 System Description…………………………….…………..38
5.2 Simulation………..………………………………..….39
Chapter 6 CONCLUSION……………………………..…….49
6.1 Conclusion.………………………………………………………..49
6.2 Recommendations……………………………….…………..50
References………………………………………………………..….51
List of Figures
Figure 2.1 (a) The original color image; (b) the converted gray scale image..……….....….6
Figure 2.2 (a) The input image; (b) the background image; (c) the moving object image
after gray-based subtraction with threshold25………...............…...………….7
Figure 2.3 Sobel operation result…….…………………………………………………….10
Figure 2.4 The nine pixels in the image……………………………………………………11
Figure 2.5 The images after morphological operation……………………………………..14
Figure 3.1 Moving object’s length and width calculation………………………………….18
Figure 3.2 Noise elimination result..…………………………………………………..........18
Figure 3.3 (a) Define eight directions; (b) the relation of eight directions and the pixel’s location…………………………………………………………………….....20
Figure 3.4 (a) An example of contour following; (b) value and point number of contour following……………………………………………………………………..21
Figure 3.5 Summary of contour following………………………………………………….22
Figure 3.6 The contour following result…………………………………………………….22
Figure 3.7 (a) Calculate the contour distance from center point; (b) the contour distance
result; (c) the smoothed distance result…………………………..…………..24
Figure 3.8 (a) The Discrete Fourier Transform result; (b) the first normalized 20 DFT
data result……………..…………………………………………………........25
Figure 4.1 Structure of the SONFIN………..…………………..……………………….…..29
Figure 4.2 Training by SONFIN…………………………………………………………….36
Figure 4.3 Recognition by SONFIN……………………………………………………..…...36
Figure 5.1 System diagram…………………………………………………………………..39
Figure 5.2 First column: some illustrating pictures of pedestrians; second column: manual
contour extraction; third column: automatic contour extraction…………………43
Figure 5.3 First column: some illustrating pictures of cars; second column: manual
contour extraction; third column: automatic contour extraction…………………44
Figure 5.4 First column: some illustrating pictures of motorcycles; second column: manual
contour extraction; third column: automatic contour extraction…………………45
Figure 5.5 First column: some illustrating pictures of dogs; second column: manual
contour extraction; third column: automatic contour extraction…………………46
Figure 5.6 Training error of SONFIN and BP (manual contour extraction)………....……….47
Figure 5.7 Training error of SONFIN and BP (automatic contour extraction)……………….47
Figure 5.8 An example to obtain feature vector by our method………………………………48
List of Tables
Table 1 Training and testing results by different networks with contours extracted
manually and automatically……………………………………………………….42

References
[1] A. Shio and J. Sklansky, “Segmentation of people in motion,” Proceedings of the IEEE Workshop on Visual Motion, pp. 325 - 332, Oct, 1991.
[2] A. G. Bors and I. Pitas, “Optical flow estimation and moving object segmentation based on median radial basis function network,” IEEE Transactions on Image Processing, Vol. 7, no. 5, pp. 693 - 702, May, 1998.
[3] A. G. Bors and I. Pitas, “Prediction and tracking of moving objects in image sequences,” IEEE Transactions on Image Processing, Vol. 9, no. 8, pp. 1441 - 1445, Aug, 2000.
[4] S. Y. Chien, S. Y. Ma and L. G. Chen, “Efficient moving object segmentation algorithm using background registration technique,” IEEE Transactions on Circuits and Systems for Video Technology, Vol. 12, no. 7 , pp. 577 — 586, July, 2002.
[5] H. Fujiyoshi and A. J. Lipton, “Real-time human motion analysis by image skeletonization,”, WACV''98. Proceedings the Fourth IEEE Workshop on Applications of Computer Vision, pp. 15 — 21, Oct, 1998.
[6] A. J. Lipton, H. Fujiyoshi and R. S. Patil, “Moving target classification and tracking from real-time video,” WACV ''98. Proceedings the Fourth IEEE Workshop on Applications of Computer Vision, pp. 8 - 14 , Oct, 1998.
[7] H. L. Tzou, Real-Time Human Detection and Tracking, Master Thesis, Tatung University, 2002.(in Chinese).
[8] P. N. Cheng, The Application of Fuzzy Inference to Automatic Detect and Identification of Intruders in Security System, Master Thesis, Tatung University, 2003.(in Chinese).
[9] I. Haritaoglu, D. Harwood and L. S. Davis, “W4 real-time surveillance of people and their activities,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 22, no. 8, pp. 809 - 830 , Aug, 2000.
[10] C. R. Wren, A. Azarbayejani, T. Darrell and A. P. Pentland, “Pfinder: Real-time tracking of the human body,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 19, no. 7, pp. 780 - 785, July, 1997.
[11] C. Kim and J. N. Hwang, “Fast and automatic video object segmentation and tracking for content-based applications,” IEEE Transactions on Circuits and Systems for Video Technology, Vol. 12, no. 2, pp. 122 — 129, Feb, 2002.
[12] C. Curio, J. Edelbrunner, T. Kalinke, C. Tzomakas and W. von Seelen, , “Walking pedestrian recognition,” IEEE Transactions on Intelligent Transportation Systems, Vol. 01, no. 3, pp. 155 — 163, Sept, 2000.
[13] I. R. Vega and S. Sarkar, “Statistical motion model based on the change of feature relationships: human gait-based recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 25, no. 10, pp. 1323 — 1328, Oct, 2003.
[14] H. Freeman, “On the encoding of arbitrary geometric configurations,” IRE Trans. Electron. Comput., vol. EC-10, pp. 260-268, June, 1961.
[15] D. Neuhoff and K. Castor, “A rate and distortion analysis of chain codes for line drawings,” IEEE Transactions on Information Theory,, Vol. 31, no. 1, Jan, 1985.
[16] C. Y. Choo and H. Freeman, “An efficient technique for compressing chain-coded line drawing images Signals,” 1992 Conference Record of The Twenty-Sixth Asilomar Conference on Systems and Computers, vol.2, pp. 717 - 720,Oct, 1992.
[17] H. Kim and K. Nam, “Object recognition of one-dof tools by a back-propagation neural net,” IEEE Transactions on Neural Networks, Vol. 6, no. 2, pp. 484 - 487, March, 1995.
[18] S. Osowski and D.D. Nghia, “Neural networks for classification of 2-D Patterns,” 5th International Conference on Signal Processing Proceedings, Vol. 3, pp. 1568 - 1571, Aug, 2000.
[19] C. T. Zahn and R. Z. Roskies, “Fourier descriptors for plane closed curves,” IEEE Trans. Comput., vol. C-21, no. 3, Mar, 1977.
[20] L. Zhao and C. E. Thorpe, “Stereo- and neural network-based pedestrian detection,” IEEE Transactions on Intelligent Transportation Systems, Vol. 01, no. 3, Sept, 2000.
[21] C. F. Juang and C. T. Lin, “An online self-constructing neural fuzzy inference network and its applications,” IEEE Transactions on Fuzzy Systems, Vol. 6, no. 1, Feb, 1998.
[22] B. Kosko, Neural Networks and Fuzzy Systems. Englewood Cliffs, NJ:Prentice-Hall, 1992.
[23] C. T. Lin, Neural Fuzzy Control Systems with Structure and Parameter Learning. New York: World Scientific, 1994.
[24] C. T. Lin and C. S. G. Lee, Neural Fuzzy Systems: A Neural-Fuzzy Synergism to Intelligent Systems. Englewood Cliffs, NJ: Prentice-Hall, 1996.