一种实时低成本压缩算法的VLSI实现

针对传统视频图像压缩算法时延长和成本高的问题,提出一种新的无损/近无损视频压缩算法。该算法由码率控制器和熵编码器组成,其中码率控制器通过对已有信息进行分析（上下文）来确定当前宏块的可用比特数,然后根据大量实验得出的高效Huffman码表,并结合位平面编码器对残差进行编码。实验结果表明,文中提出的视频图像压缩算法能够工作在300 MHz,吞吐量最差为1.3 pixel/cycle,同时仅用一块120＊720的SRAM来存储上一行像素值,因此很好地解决了时延和成本问题。

Realization of VLSI of an Real-time and Low Cost Compression Algorithm

In view of the problem of time delay and high cost for traditional video image compression algorithm,a novel lossless/near-lossless video compression algorithm is proposed. This algorithm consists of rate controller and entropy encoder,the rate controller can de-termine the bits of current block by analyzing the information,then according to the Huffman code table obtained by plenty of experi-ments,the residual is encoded combined with bit-plane encoder. The experimental results illustrate that the proposed algorithm can work in 300 MHz and the worst throughput is 1. 3 pixel/cycle,while only a 120＊720 SRAM is used to store the top-line pixel,thus the prob-lem of time delay and high cost can be solved properly.