An Improved SPIHT Algorithm for Lossy Image Coding
Baolin Zhou · 2021
The raw image data occupies a large storage space and communication bandwidth. Compression of the image data saves storage space and transmission bandwidth by removing the redundancy. In many image compression algorithms, wavelet transform can decompose the image with multi-level resolution, and can perform embedded coding. The author studies the set partitioning in hierarchical trees (SPIHT) algorithm based on wavelet transform and proposed an improved SPIHT algorithm (NE-SPIHT). First, the initialization of LIS in the SPIHT coding scheme is improved, The LIP and LIS are initialized to the nodes with no descendants in the lowest frequency subband. Secondly, when the nodes in the LIS chain table are judged as significant, its four child nodes are regarded as one whole for judgment, reducing the output bits. Experimental results demonstrate that the NE-SPIHT algorithm outperform SPIHT in terms of PSNR and SSIM, so applying NE-SPIHT method to the practical image compression can achieve a high compression efficiency.