Reference frame data compression method for H.264/AVC

Abstract

In this paper a novel reference frame data compression method for the H.264/AVC is proposed. Proposed method introduce a new function including the DAP (Differential of Adjacent Pixel) method and Huffman coding tool between the frame memory and the image coding engine. Simulation results show that this method could achieve not only a stable compression rate but also realize a lossless compression without additional distortion to the coding process. Furthermore, proposed method will not induce further delay during the memory access period. Proposed method is hardware-friendly and could be implemented by only about 7,000 gates.