An Approach of Automatic Motion Transverse Extension

ABSTRACT We present a concept and technique of human motion reuse based on motion extension, which contains lengthways extension and transverse extension. Motion extension can generate more abundant new motion from finite motion data, which provides uniform solution for motion forecast, motion repair and group motion. Lengthways extension extends human motion along time axis and generates new motion containing more frames, and transverse extension extends human motion along characters, which generates multiple character group motion from single character motion. Present an approach of automatic motion transverse extension, which can generate more characters motion from single character motion. Experiments show that motion transverse extension can advance the reuse of sample motion. Keywords: virtual human, motion, transverse extension, reuse