Classification of Human Shapes Using Angular Radial Transform
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This paper deals with a new classification algorithm of human shapes using the region-based shape descriptor and mean shift clustering. The main goal is to identify that the extracted binary object regions are human shapes. This paper combines the angular radial transform (ART) with mean shift clustering, which is a new approach to human shape classification. The ART, a region-based shape descriptor in MPEG-7, is applied to model the binary human shapes. We exploit 3 radial and 12 angular frequencies for modeling human shapes. The 36-D ART vectors for human shapes are trained using mean shift clustering, and several representative ART vectors are selected for multi-mode distribution of human shapes in 36-D ART vector space. The human shapes are identified by the distance between the modeled representative vectors and ART vector of extracted object region. The distance thresholds are statistically optimized in the Training step. The ART vectors of human shapes are learned using thousands of illustration images and real human objects extracted by background subtraction. Experiments are performed on various illustration images and video frames combined with background subtraction. The experimental results show that the proposed algorithm is robust and efficient for human object recognition. The proposed method is invariant and stable to object rotation and scales. It is expected that the proposed algorithm of human shapes recognition is useful for object recognition, video surveillance, vehicular pedestrian detection, and so on.
Abstract
This paper deals with a new classification algorithm of human shapes using the region-based shape descriptor and mean shift clustering. The main goal is to identify that the extracted binary object regions are human shapes. This paper combines the angular radial transform (ART) with mean shift clustering, which is a new approach to human shape classification. The ART, a region-based shape descriptor in MPEG-7, is applied to model the binary human shapes. We exploit 3 radial and 12 angular frequencies for modeling human shapes. The 36-D ART vectors for human shapes are trained using mean shift clustering, and several representative ART vectors are selected for multi-mode distribution of human shapes in 36-D ART vector space. The human shapes are identified by the distance between the modeled representative vectors and ART vector of extracted object region. The distance thresholds are statistically optimized in the Training step. The ART vectors of human shapes are learned using thousands of illustration images and real human objects extracted by background subtraction. Experiments are performed on various illustration images and video frames combined with background subtraction. The experimental results show that the proposed algorithm is robust and efficient for human object recognition. The proposed method is invariant and stable to object rotation and scales. It is expected that the proposed algorithm of human shapes recognition is useful for object recognition, video surveillance, vehicular pedestrian detection, and so on.
참고문헌 (33)
- M. Valera, "Intelligent distributed surveillance systems: A review", IEEE Proc. Vis. Image, Signal Process, 152(2), pp.192~204 (2005)
- A. Hampapur, "Smart video surveillance: Exploring the concept of multiscale spatiotemporal tracking", IEEE Signal Process. Mag, 22(2), pp.38~51 (2005)
- W. Hu, "A survey on visual surveillance of object motion and behaviors", IEEE Trans. Syst., Man, Cybern. C, Appl. Rev., 34(3), pp.334~352 (2004)
- Massimo Piccardi, "Background subtraction techniques: a review", IEEE International Conf. on Systems, Man and Cybernetics, pp.3099~3104 (2004)
- Thanarat Horprasert, "A statistical approach for real-time robust background subtraction and shadow detection", Proc. of IEEE ICCV Frame-Rate Workshop (1999)
- P. Kaew, "An improved adaptive background mixture model for realtime tracking with shadow detection", Proc. of European Workshop on Advanced Video Based Surveillance Systems (2001)
- M. Seki, "Background subtraction based on co-occurrence of image variations", Proc. of CVPR, 2, pp.65~72 (2003)
- S. Shafer, "Using color to separate reflectance components", Color Research and Applications, 10, pp.210~218 (1985)
- M. Bober, "MPEG-7 visual shape descriptors", IEEE Trans. Circuits Syst. Video Technol, 1(6) (2001)
- Julien Ricard, "Generalizations of Angular Radial Transform for 2D and 3D Shape Retrieval", Pattern Recognition Letters, 26, pp.2174~2186 (2005)
- S. K. Hwang, "Fast and efficient method for computing ART", IEEE Trans. Image Proc, 15(1), pp.112~117 (2006)
- M.-H. Yang, "Detecting faces in images: A survey", IEEE Trans. on pattern Analysis and Machine Intelligence, 24(1), pp.34~58 (2002)
- R.-L. Hsu, "Face detection in color images", IEEE Trans. Pattern Analysis and Machine Intelligence, vol, 24, pp.696~706 (2002)
- P. C. Hsiehand, "Similarity retrieval of shape images based on database classification", Jour. of Visual Comm. and Image Repre. (2006)
- D. Comaniciu, "Mean shift: A robust approach toward feature space analysis", IEEE Trans. Pattern Analysis and Machine Intelligence, 24(5), pp.603~619 (2002)
- Chung-Chen Chen, "Cooperative mapping of multiple PTZ cameras in automated surveillance systems", Proc. of IEEE CVPR (2009)
- Piotr Doll, "Pedestrian detection: A benchmark", Proc. of IEEE CVPR (2009)
- Branislav Micusik, "Simultaneous surveillance camera calibration and foot-head homology estimation from human detections", Proc. of IEEE CVPR (2010)
- Maria Sagrebin, "Real-Time Moving Object Detection for Video Surveillance", Proc. of Advanced Video and Signal Based Surveillance (2009)
- Ki-yeol Eom, "Fast Object Tracking in Intelligent Surveillance System", Proc. of Computational Science and Its Applications (2009)
- Parisa Darvish, "Human Tracking by IP PTZ Camera Control in the Context of Video Surveillance", Proc. of International Conference on Image Analysis and Recognition (2009)
- Yiliang Xu, "Real-Time Object Tracking and Classification Using a Static Camera", Proc. of Int l Conf. on Intelligent Robots and Systems (2009)
- M. Isard, "Condensation-conditional density propagation for visual tracking", International Journal of Computer Vision, vol, 29(1), pp.5~28 (1998)
- R. C. Gonzalez, "Digital Image Processing" (1994), Addison Wesley
- G. Borgefors, "Hierarchical chamfer matching: a parametric edge matching akgorithm", IEEE Trans. Pattern Analysis and Machine Intelligence, 10(6), pp.849~865 (1988)
- F. Mokhtarian, "Silhouette-based isolated object recognition through curvature scale space", IEEE Trans. Pattern Analysis and Machine Intelligence, 17(5), pp.539~544 (1995)
- M. K. Hu, "Visual pattern recognition by moments invariants", RE Trans. Information Theory, IT-8, pp.179~187 (1962)
- J. Fang, "Human face detection using angular radial transform and support vector machines", Proc. of ICIP (2003)
- N. Dalal, "Histograms of oriented gradients for human detection", Proc. of CVPR, 1, pp.886~893 (2005)
- Ronald Poppe, "Evaluating example-based pose estimation: Experiments on the HumanEva sets", Proc. of CVPR (2007)
- 손효찬, "배경차분과HOG의결합을 통한 실시간 보행자 탐지 기법", 영상처리 및 이해에 관한 워크샵 논문집, pp.91~93 (2011)
- W. R. Schwartz, "Human detection using partial least squares analysis", Proc. of ICCV (2009)
- J. Shotton, "Real-time human pose recognition in parts from single depth images", Proc. of CVPR (2011)