Lossless Image Compression Using Dynamic Block Encoding
Fangxun Shu, Xuelin Yang · 2020
In this paper, we present a simple and effective algorithm of lossless image compression to save coding space. Dynamic block encoding is applied, to reduce the coding bit via the reduction of the image spatial correlation. The original image is divided into uniform blocks with a proper size, to achieve a good intra-block correlation, then a reference rule is applied to decorrelate each block and squeeze the extra coding bit of each block, resulting in effective compression. Compared with the traditional unified encoding, the proposed algorithm optimizes the coding space via intra-block correlation, and the iteration compression can be performed further to squeeze inter-block correlation fully. Compression ratios of 1.807 and 1.857 are achieved for a single and three iterations, respectively, with a block size of 4. Due to the simple and efficient reference rule, the proposed algorithm also avoids complex mathematical computation in the traditional image compression methods.