부산대학교 도서관

In this paper, a novel lossless and near-lossless RAW image compression algorithm based on JPEG-LS is proposed. To achieve high performance and low complexity, the algorithm reduce the context module, merge the run mode with the regular mode and propose an adaptive Golomb-Rice coding technique. The algorithm mainly consists of four parts : pixel structure conversion, pixel prediction, run mode module and adaptive Golomb-Rice coding module. Compared to JPEG-LS, this algorithm achieves comparable performance to JPEG-LS while reducing complexity. Additionally, it allows users to adjust parameters to achieve lossless and near-lossless compression supporting random access and non-random access.