Gabor feature-based tire tread patterns recognition by support vector machine 陳靜宜、黃登淵
E-mail: [email protected]
ABSTRACT
Tire tread patterns are broadly used in the investigation of traffic accidents for recognizing the responsibility of car drivers. They are quite useful especially for a hit-and-run accident by identifying a given tread pattern to match existing tires to further reduce the scope of investigation. The above information is of importance in forensic evidence for law enforcement agencies. However, most pattern matching processes are manually operated, which are labor-intensive and require visual identifications of extensive database of tire tread patterns. In this thesis, we propose to automate the matching process of tire tread patterns by creating effective features representation, extraction and classification, and then to locate candidate matches from the database of existing tread pattern images. In the proposed algorithm, input tire images are first preprocessed by binarization as well as fast 8-connected component labeling method to enhance the features on tire surface. The grooves in tire surface are salient important features for our tire matching system. We detect the tire tread patterns of being grooved or wavy and use this feature to train several SVM classifiers.
The features of tire tread patterns are then represented by Gabor wavelets, and feature extraction is further carried out by principal component analysis (PCA). The matching processes are achieved by the classifiers of SVM, Euclidean distance and cosine distance.
Result shows that the recognition rate of 83% for tire images can be obtained by the SVM classifier when 15 tire tread patterns are used.
Keywords : tire tread patterns recognition、Gabor feature、PCA、SVM Table of Contents
封面內頁 簽名頁 授權書．．．．．．．．．．．．．．．．．．．．．．．．．iii 中文摘要．．．．．．．．．．．．
．．．．．．．．．．．．iv 英文摘要．．．．．．．．．．．．．．．．．．．．．．．．v 誌謝．．．．．．．．．
．．．．．．．．．．．．．．．．．vi 目錄．．．．．．．．．．．．．．．．．．．．．．．．．．vii 圖目錄．．．
．．．．．．．．．．．．．．．．．．．．．．x 表目錄．．．．．．．．．．．．．．．．．．．．．．．．．xii 第 一章 緒論 1.1研究背景．．．．．．．．．．．．．．．．．．1 1.2文獻回顧與探討．．．．．．．．．．．．．．．2 1.3研究方法．．．．．．．．．．．．．．．．．．4 1.4研究結果．．．．．．．．．．．．．．．．．．4 1.5本文架構
．．．．．．．．．．．．．．．．．．5 第二章 車胎紋路辨識系統 2.1前言．．．．．．．．．．．．．．．．．．．
．6 2.2自建車胎紋路資料庫．．．．．．．．．．．．．7 2.2.1 輪胎規格型號．．．．．．．．．．．．．．．9 2.2.2 垂 直與水平鏡像．．．．．．．．．．．．．．12 2.3二值影像之應用．．．．．．．．．．．．．．．13 2.3.1 影像二值化
．．．．．．．．．．．．．．．．13 2.3.2 連通區域—8連通法．．．．．．．．．．．．14 2.3.3 影像形態學．．．．
．．．．．．．．．．．．18 2.3.4 偵測溝痕數．．．．．．．．．．．．．．．．19 2.4影像辨識．．．．．．．．．．
．．．．．．．．20 2.5相關軟硬體之規格．．．．．．．．．．．．．．20 第三章 車胎紋路的特徵擷取與辨識方法 3.1 前言．．．．．．．．．．．．．．．．．．．．21 3.2Gabor Filter原理．．．．．．．．．．．．．22 3.2.1 Gabor Feature．．．．．．．．．．．．．．24 3.3主分量分析(PCA)理論基礎．．．．．．．．．．26 3.3.1 傳統主分量分析方 法(PCA) ．．．．．．．．．29 3.4車胎紋路辨識分類器．．．．．．．．．．．．．31 3.4.1 歐式距離分類器(Euclidean Distance Classifier)．31 3.4.2 餘弦距離分類器(Cosine Distance Classifier)．32 3.4.3 支持向量機分類器(Support Vector Machine Classifier；SVM Classifier)．．．32 第四章 支持向量機(SVM) 4.1前言．．．．．．．．．．．．．．．．．．．．34 4.2線 性可分離．．．．．．．．．．．．．．．．．35 4.3線性不可分離．．．．．．．．．．．．．．．．38 4.4非線性可分 離．．．．．．．．．．．．．．．．40 4.5支持向量機之核函數選擇與參數設定．．．．．．43 第五章 車胎紋路辨識系 統流程與實驗結果 5.1車胎紋路辨識系統流程．．．．．．．．．．．．45 5.2實驗結果與討論．．．．．．．．．．．．
．．．48 第六章 結論與未來研究方向 6.1結論．．．．．．．．．．．．．．．．．．．．53 6.2未來研究方向．．．．
．．．．．．．．．．．．53 參考文獻．．．．．．．．．．．．．．．．．．．．．．．54 REFERENCES
[1]王瑩瑋、林昭男,剎車胎痕自動偵測模式之建立,交通學報,第四卷,第二期,頁15-27,民國九十三年十二月.
[2]林昭男, “以影像處理技術為基礎之胎痕識別系統,”大葉大學車輛工程研究所碩士論文, 2004.
[3]Y.W. Wang, “A Tire-mark Identification Scheme for Suspected Vehicle Detection in hit and run Accident,” Journal of the Eastern Asia Society for Transportation Studies, Vol.6, pp. 3441-3452, 2005.
[4]Y.W. Wang, “A Tire Mark Localization Method for Forensic Image Analysis,” Journal of the Eastern Asia Society for Transportation Studies, Vol. 7, pp. 2881-2890, 2007.
[5]M.Y. Kalinkin, Usanov, D.A., and A.V. Skripal, “Recognition of tire pattern on its trace by computer video technology,” Proceedings of SPIE, Vol.5772, pp.199-202, 2004.
[6]M.J. Thali, M. Braun, W. Br?卲chweiler, and R.Dirnhofer, “Matching tire tracks on the head using forensic photogrammetry,” Forensic Science International, Vol. 113, pp 281-287, September 2000.
[7]W. Br?卲chweiler, M. Braun, R. Dirnhofer, and M.J. Thali, “Analysis of patterned injuries and iniury-causing instruments with forensic 3D/CAD supported photogrammetry (FPHG): an instruction manual for the documentation process,” Forensic Science International, Vol.132, pp 130-138, 2003 [8]U. Buck, N. Albertini, S. Naether, and M. J. Thali, “3D documentation of footwear impressions and tyre tracks in snow with high resolution optical surface scanning,” Forensic Science International, Vol. 171, pp 157-164, Sep. 2007.
[9]S.W. Jung, S.W. Bae, and G.T. Park, “A design scheme for a hierarchical fuzzy pattern matching classifier and its application to the tire tread,
” Fuzzy Sets and Systems, Vol. 65, pp 311-322, Aug. 1994.
[10]D. Colbry, D. Cherba, and J. Luchini, “Pattern recognition for classification and matching of car tires,” Tire Science and Technology, Vol.
33, pp 2-17, Jan. 2005.
[11]P. McDonald, “Tire Imprint Evidence,” Elsevier series in Practical aspects of criminal and forensic Investigation, New York, 1989.
[12]P. Moreno, A. Bernardino, J. S. Victor, “Gabor parameter selection for local feature detection,” Proceedings of IbPRIA2005, Lecture Notes in Computer Science 3522, pp. 11-19, 2005.
[13]H. Y. Chen, C. L. Huang, and C. M. Fu, “Hybrid-boost learning for multi-pose face detection and facial expression recognition,” Pattern Recognition, Vol. 41, pp.1173-1185, 2008.
[14]A. M. Martnez, and A. C. Kak, “PCA versus LDA,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 23, No. 2, pp.
228-233, 2001.
[15]P. N. Belhumeur, J. P. Hespanha, and D. J. Kriegman, “Eigenfaces vs. Fisherfaces: recognition using class specific linear projection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 19, No. 7, pp. 711-720, July 1997.
[16]J. Wang, K. N. Plataniotis, and A. N. Venetsanopoulos, “Selecting discriminate eigenfaces for face recognition,” Pattern Recognition Letters, Vol. 26, pp. 1470-1482, 2005.
[17]J. Wang, K. N. Plataniotis, and A. N. Venetsanopoulos, “Selecting discriminate eigenfaces for face recognition,” Pattern Recognition Letters, Vol. 26, pp. 1470-1482, 2005.
[18]Z. Shaoyan and Q. Hong, “Face recognition with support vector machine,” In: Proceedings of IEEE International Conference on Robotics, Intelligent Systems and Signal Processing, Vol. 2, pp. 726-730, Changsha, China, 2003.
[19]L. Bing, Z. Yun, and P. Y. Hong, “Face recognition based on wavelet transform and svm,” IEEE International Conference on Information Acquisition, pp. 373-377, Hong Kong and Macau, China, 2005.
[20]G. Guodong, S. Z. Li, and C. Kapluk, “Face recognition by support vector machines,” Proceedings of IEEE Conference on International Automatic Face and Gesture Recognition, pp. 196-201, Grenoble, France, 2000.
[21]J. Qin, and Z. S. He, “A SVM face recognition method based on Gabor-feature key points,” Proceedings of the Fourth International Conference on Machine Learning and Cybernetics, pp. 5144-5149, Guangzhou, China, 2005.
[22]K. Jonsson, J. Kittler, Y. P. Li, and J. Matas, “Support vector machines for face authentication,” Proceedings of British Machine Vision Conference Nottingham, pp. 543-553, Nottingham, 1999.
[23]M. Safari, M. T. Harandi, and B. N. Araabi, “A SVM-based method for face recognition using a wavelet PCA representation of faces,” In:
International Conference on Image Processing, Vol.2, pp. 853-856, Singapore, 2004.
[24]V. N. Vapnik, “Statistical Learning Theory,” John Wiley & Sons, Inc., New York, 1998.
