Using 2-D and 3-D Ellipsoid Fitting for Head and Body Segmentation and Head Tracking

In this paper, a novel procedure is presented for segmenting a general 3-D wireframe model obtained from a head and shoulders multiview sequence. The procedure consists of two steps. In the rst step, two ellipses corresponding to the head and the body of the person are identi ed based on ellipse tting of the outline of the person in each image. The tting is based on a direct least squares method using the constraint that forces a general conic to be an ellipse. In order to achieve head/body segmentation, a K-means algorithm is used to minimize the tting error between the points and the two ellipsoids. In the second step, an 3-D ellipsoid model corresponding to the head of the person is identi ed using an extension of the above method. Thus a coarse approximation of the initial wireframe requiring minimal bitrate can be obtained. The proposed technique is fast and yields very accurate results.