A hybrid scheme using improved SPIHT and Huffman coding for lossless image compression

In order to improve the coding performance of image lossless compression, a hybrid lossless compression scheme based on modified Set Partitioning in Hierarchical Trees (SPIHT) and Huffman coding is proposed. The lower magnitude bit-planes, whose corresponding magnitude thresholds are less than 8, and the sign bit-plane are scanned node by node in the fixed order, and the bit sequence obtained by scanning the bit-planes is further processed by Huffman coder to remove the statistical redundancy. The other bit-planes, namely the higher magnitude bit-planes, are sorted by the modified SPIHT partitioning structure, which could improves the coding efficiency of sorting pass by means of grouping more nodes together. Experimental results show the proposed algorithm has reduced its bit-rate by 0.13bit/pixel-0.35bit/pixel for gray images and by 0.72bit/pixel-1.25bit/pixel for color images as compared with those of the original SPIHT algorithm. Furthermore, the modified scheme retains the progressive transmission property of SPIHT coder. It is concluded that the proposed method efficiently reduces the redundancy existing in the original SPIHT coder and is fit for lossless image compression.