Multiple curvature based approach to human upper body parts detection with connected ellipse model fine-tuning

In this paper, we discuss an effective method for detecting human upper body parts from a 2D image silhouette using curvature analysis and ellipse fitting. First we smooth the silhouette so that we can determine just the global features: the head, hands and armpits. Next we reduce the smoothing to detect the local features of the neck and elbows. We model the human upper body by multiple connected ellipses. Thus we segment the body by the extracted features. Ellipses are fitted to each segment. Lastly, we apply a non-linear least square method to minimize the differences between the connected ellipse model and the edge of the silhouette.

[1]  Josef Kittler,et al.  A New Approach to Computation of Curvature Scale Space Image for Shape Similarity Retrieval , 1997, ICIAP.

[2]  Pascal Fua,et al.  Tracking articulated bodies using Generalized Expectation Maximization , 2008, 2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops.

[3]  Chin-Chen Chang,et al.  Hand pose recognition using curvature scale space , 2002, Object recognition supported by user interaction for service robots.

[4]  Gary R. Bradski,et al.  Motion segmentation and pose recognition with motion history gradients , 2000, Proceedings Fifth IEEE Workshop on Applications of Computer Vision.

[5]  Dariu Gavrila,et al.  A Bayesian, Exemplar-Based Approach to Hierarchical Shape Matching , 2007, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  Vincent Lepetit,et al.  Human body pose detection using Bayesian spatio-temporal templates , 2006, Comput. Vis. Image Underst..

[7]  Ferran Marqués,et al.  Silhouette-based probabilistic 2D human motion estimation for real-time applications , 2005, IEEE International Conference on Image Processing 2005.

[8]  P. Fua,et al.  Tracking Of Hand ’ s Posture And Gesture , 2004 .

[9]  Gary R. Bradski,et al.  Motion segmentation and pose recognition with motion history gradients , 2002, Machine Vision and Applications.

[10]  W. Eric L. Grimson,et al.  Adaptive background mixture models for real-time tracking , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).