JPEG自适应量化参数选择算法

Research on Quantization Distortion Estimation Algorithm of Jpeg

量化作为图像压缩中主要的失真来源之一,对量化过程进行优化可提升编码质量。为提高图像的压缩比和压缩质量,结合信息论中的注水理论及反注水模型,提出了利用自适应反注水量化参数构建的方法,采用新的量化表替代JPEG中的量化表,通过JPEG的编码算法对多幅不同的灰度图像进行了压缩仿真验证,同时与JPEG压缩作对比分析。实验结果表明,相同压缩比下,提出的自适应量化表的PSNR与JPEG压缩方法相比平均提高了0.3322dB,提出的反注水的自适应算法是一种有效、压缩效果更佳的量化方法。

As one of the main sources of distortion in image compression, the optimization of quantization process can improve the coding quality. In order to improve the compression ratio and quality of the image, this paper proposes a method to construct the self-adaptive anti water injection quantization parameters by combining the water injection theory and the anti water injection model in the information theory. A new quantization table was used to replace the quantization table in JPEG. Through the coding algorithm of JPEG,the compression simulation experiment of several different gray-scale images was carried out and compared with JPEG compression. The experimental results show that the PSNR of the self-adaptive quantization table proposed in this paper is 0.3322 db higher than that of JPEG compression method on average under the same compression ratio, which shows that the self-adaptive algorithm based on anti water injection proposed in this paper is an effective quantization method with better compression effect.