View-Point Insensitive Human Pose Recognition using Neural Network

Authors: Sanghyeok Oh, Yunli Lee, Kwangjin Hong, Kirak Kim, Keechul Jung

Abstract:

This paper proposes view-point insensitive human pose recognition system using neural network. Recognition system consists of silhouette image capturing module, data driven database, and neural network. The advantages of our system are first, it is possible to capture multiple view-point silhouette images of 3D human model automatically. This automatic capture module is helpful to reduce time consuming task of database construction. Second, we develop huge feature database to offer view-point insensitivity at pose recognition. Third, we use neural network to recognize human pose from multiple-view because every pose from each model have similar feature patterns, even though each model has different appearance and view-point. To construct database, we need to create 3D human model using 3D manipulate tools. Contour shape is used to convert silhouette image to feature vector of 12 degree. This extraction task is processed semi-automatically, which benefits in that capturing images and converting to silhouette images from the real capturing environment is needless. We demonstrate the effectiveness of our approach with experiments on virtual environment.

Keywords: Computer vision, neural network, pose recognition, view-point insensitive.

Digital Object Identifier (DOI): doi.org/10.5281/zenodo.1332444

References:

[1] Y. Sagawa, M. Shimosaka, T. Mori and T. Sato, "Fast Online Human Pose Estimation via 3D Voxel Data", Intelligent Robots and Systems, pp. 1034-1040, 2007.
[2] Catherine A., Xingtai Q., Arash M., Maurice., "A novel approach for recognition of human actions", Machine Vision and Applications, 2008, pp. 27-34, 2008.
[3] M. Voit, K. Nickel, R. Stiefelhagen, "Neural Network-Based Head Pose Estimation and Multi-view Fusion", LNCS 4122, pp. 291-298, 2007.
[4] C. Yuan, H. Niemann, "Neural networks for the recognition and pose estimation of 3D objects from a single 2D perspective view", Image and Vision Computing 19, pp. 585-592, 2001.
[5] URL downloads 3D models : http://www.turbosquid.com
[6] A. Agarwal, B. Triggs, "Human Pose from Silhouettes by Relevance Vector Regression", CVPR, vol2., pp882-888, 2004.
[7] R. Rosales, S. Sclaroff, "Learning and synthesizing human body motion and posture", IEEE, 2000.
[8] F. Lb, R. Nevatia, "Single View Human Action Recognition using Key Pose Matching and Viterbi Path Searching", CVPR, 2007.
[9] H. Yu, G. Sun, W. Song, X. Li, "Human Motion Recognition Based on Neural Network", IEEE, Vol. 2, pp. 982, 2005.
[10] M. Voit, K. Nicket, R. Stiefelhagen, "Multi-view Head Pose Estimation using Neural Networks", Computer and Robot Vision, pp. 347-352, 2005.