Motion-compensated wavelet transform coding
Fréderic Dufaux, F. Moscheni, Markus Schütz · 1994
The wavelet transform has proved its efficiency in removing spatial redundancy. In the framework of hybrid coding schemes, this paper studies the way to combine the wavelet transform and a block matching motion estimation technique. Different coding schemes are presented which differ through the data on which the motion estimation and motion compensation are respectively applied. Simulations results are presented which compare the performances of the different schemes. 1 Introduction Related to the emergence of digital television, there has been lately much interest in video coding. For this purpose, hybrid coding schemes have proved their efficiency. They rely on two basic techniques: transform coding for the reduction of the spatial redundancies and motion estimation and compensation for the reduction of the temporal redundancies. The MPEG1 [1] and MPEG2 [2] standards, based on the Discrete Cosine Transform (DCT) and block matching motion estimation, are two examples of such schemes...