Compression coding of moving contours on spatio-temporal 2-D plane

A compression coding technique for moving contours is proposed. A set of moving contours is first resampled uniformly to obtain the same number of sampled points on each contour and then the n-th sampled point on the m-th contour (x_m(n)y_m(n)) is arranged on an m-n spatio temporal plane. Since the horizontal and vertical correlations of x_m(n) and y_m(n) are considerably high, they can be compressed by using existing image coding techniques. Wavelet-based and DCT-based techniques are utilized and several experimental results are shown. The CAE method is also tested for the same moving contours. As a result, it is confirmed that the wavelet-based compression technique on the spatio-temporal 2-D plane gives the best coding efficiency in the lower bit rate region