Adaptive Pipeline Hardware Architecture Design and Implementation for Image Lossless Compression/Decompression Based on JPEG-LS

Abstract

Lossless compression of medical images can reduce data size, save storage and transmission costs, and cope with the challenge of increasing resolution demand for medical images. This paper proposes an adaptive pipeline hardware architecture for lossless compression/decompression of medical images based on joint photographic experts group-lossless (JPEG-LS), which can achieve efficient image processing on field programmable gate array (FPGA). In addition, by adding a pipeline pause mechanism, the problem of coding errors caused by parameter updates is solved. On the one hand, the binary search method is used to optimize the calculation process of the Golomb coding parameter k in the pipeline, which reduces the parameter update delay. On the other hand, the construction process of the context causal template is optimized, which effectively improves the throughput. Based on the parallel block compression architecture, a hardware module for alpha red green blue (ARGB) image block compression/decompression with customizable block size is implemented, which improves the image compression/decompression throughput and reduces the compression and decompression delay. This paper implements the hardware architecture on Zynq7000 FPGA and evaluates its performance. Compared with existing schemes, the proposed architecture uses less resources and achieves higher compression ratio and clock frequency.