Color image compression algorithm combining the human contrast perception characteristics

Abstract

In order to make the compressed images satisfy more with the human visual perception char acteristics, a color image compression algorithm combined the human visual contrast sensitivity characteristics is proposed. Firstly, combining the characteristics of human visual contrast sensitivity and spectrum coefficients in transform domain of image, a method for calculating the angular frequencies of the images is proposed, and the calculating method of minimum error threshold of human eye detecting images is proposed according to the angular frequencies. Then taking the threshold as quantization step size, a method is proposed to quantize spectrum coefficients in transform domain of image. Finally Huffman coding algorithm is adopted to carry out encoding and decoding to realize image compression. And the simulation experiments for three color images is carried out. The results show that the average compression ratio, PSNR and SSIM of three color images and each component of the image are increased by 10. 480 7%, 6. 987 9% and 2. 649 4% respectively, compared with the technology of JPEG. It indicates that the image compression algorithm combined the human visual characteristics is better and more practical.