改进的快速算术编码及其在图像编码中的应用

An Improved Fast Arithmetic Coding and Its Application in Image Coding

为了快速地进行图像编解码,提出了一种改进的快速算术编码方法,该方法是先通过编码区间宽度值与阈值的比较,并以字节为单位进行重归一化,同时输出以字节为单位的编码流;然后通过对当前输出字节的内容进行判断及处理,以便在不额外增加码流的情况下更加有效地处理编码过程中的进位传播的问题,从而简化了编解码过程。实验表明,在使用相同的二进制索引树概率模型时,该方法比CACM87及Jiang提出的方法具有更快的执行速度。最后,将该方法用于以2×2系数块为单位的SPIHT图像编码中,以降低位平面编解码过程的复杂度。

An improved fast arithmetic coding is proposed in order to encode and decode image quickly. It does renormalization in byte style by comparing the width of coding range with threshold, and the output codes are in byte style. In addition, it deals with the carry propagation problem more effectively without the additional output codes by judging the current output byte. So the proposed method significantly simplifies the process of encoding and decoding. The experimental results show that the proposed method runs faster than CACM87 and the method proposed by Jiang. We applied the proposed method to a 2 × 2 block based SPIHT image coding and the results show that the method can simplify the process of bitplane encoding and decoding.