A parallel hybrid video coding method based on noncausal prediction with multimode

Abstract

This paper addresses the parallel computing problem of hybrid video coding method. In particular, we proposed a new adaptive hybrid video coding method of I-Frame based on noncausal prediction which has better parallel performance than traditional causal prediction. However, there is an inherent problem of noncausal prediction: the error will be expanding when decoded. In order to solve this problem, feedback quantization has also been applied. Another character of this method is that the transform and scan order can be updated according to the input images and quantized step. The simulation results show that the proposed method is 0.4-5dB superior to H.264 High complexity profile which uses RD technology. © 2011 Springer-Verlag.