Wavelet transform coding of image sequence

In MPEG, the image sequence compression is achieved by motion compensation, transformation, quantization and entropy coding. In this paper, we follow the same path by using wavelet based modules in the coding chain. The prediction error signal is obtained by using variable block size motion compensation (VBSMC). Great attention has been paid to smooth the motion vector field, which helps to increase the coding efficiency. After VBSMC, discrete wavelet transform (DWT) is then applied to the prediction error signal to obtain wavelet coefficients. Some new schemes are introduced. First, a new and effective hybrid coder using quadtree and lattice vector quantization (QTLVQ) is introduced, which can efficiently code the wavelet coefficients of the motion compensated prediction error signal. Second, a new rate allocation scheme different from traditional ones is introduced, which can flexibly control the coder´s output rate and the duality of reconstructed video signal