改进的SPIHT静止图像压缩编码算法

Improved SPIHT still image compression algorithm

SPIHT算法是压缩编码效率很高的静止图像压缩编码算法。针对原算法没有从最佳率失真的角度出发来选择初始量化门限这一不足之处 ,提出了一种改进的 SPIHT算法。通过理论计算和试验分析 ,得到初始量化门限 T0 和编码失真的关系 ,并导出搜索最优初始量化门限 T0 的简单迭代方法。改进的 SPIHT算法能根据输入图像的特性和给定的编码输出码率自适应地选择最优初始量化门限 T0 。在相同输出码率的条件下 ,改进的 SPIHT算法比原算法峰值信噪比提高最多达 0 .6 d

The SPIHT algorithm is one of the most efficient still image compression algorithms. However, a shortcoming is that the initial quantization threshold T 0 of SPIHT is not considered in the optimal rate distortion sense. An improved SPIHT algorithm was proposed to choose an appropriate threshold. The relationship between the initial quantization threshold T 0 and distortion was used to develop a search method for the optimal initial quantization threshold T 0 . The improved SPIHT algorithm can choose the optimal initial quantization threshold T 0 according to the characteristics of the input image and the given output bit rate. For the same output bit rate, the improved SPIHT provides PSNR performance up to 0.6dB better than the original.