整数小波的有损与无损图像压缩

提出了基于整数小波变换的有损与无损图像编码方案。有损编码采用无链表的SPIHT零树编码算法,它不同于SPIHT和LZC算法的零树分割策略和状态比特表结构。该算法所需的存储空间小,有损压缩性能高,易于硬件的实现。无损编码根据不同子带小波系数的分布特性,采用带间预测编码,不同方向的子带采用不同的预测方式,预测误差采用霍夫曼编码。测试结果证明,基于整数小波的无链表有损压缩方案不仅优于LZC,接近于SPIHT,而且易于硬件的实现。

Image coding with lossy and lossless compression on integer wavelet transform

Image loss coding algorithm and lossless coding algorithm based on integer wavelet transform were proposed. In the former, SPIHT without list was used, that was different from SPIHT and LZC in zerotree segmentation and bit list of state, in addition, being with little memory space. In the latter, different predictions on each subband were adopted. Basing on different distributing property of wavelet coefficients, prediction errors were encoded by Huffman. The experimental results indicate that the proposed algorithm based on integer wavelet are superior to LZC and close to SPIHT, in addition, being beneficial to realizing of compression system on hardware.