Video object extraction and shape coding using B-splines

This work presents a system for semi-automatic video object extraction and efficient shape coding using cubic B-splines. In the framework of object-based layered coding of image sequences, shape information is essential for content-based access to video objects, and its efficient encoding needs to be investigated. We present a rate and distortion controlled algorithm for video object shape approximation by variable number of cubic B-spline segments and motion compensated inter-frame coding of B-spline control points. Video object masks are generated using a semi-automatic tracking algorithm. Given an outline of the object in the first frame of the sequence, color and shape models for video object are generated. The adjustment of object's shape and position in the following frames is based on motion estimation and local color information in the neighborhood of the object's boundary. Ratedistortion efficiency of the proposed shape coding algorithm is compared to MPEG-4 context arithmetic encoding.