为了能进行实时有效的计算机桌面图像传输,提出了一种基于块分类的计算机桌面图像混合压缩算法,该算法根据桌面图像中文本、图形和图像各部分在色彩和纹理上的不同特性,将块分为文本/图形块、混合块和图像块3类,其中,文本/图形块采用调色板编码、Hextile编码和LZW编码相结合的压缩方式,图像块采用JPEG有损压缩,混合块则采用了两者的混合算法,实验结果表明,该算法能够准确地分割出文本/图形类和图像类信息,与其他混合编码方法相比,重建图像的峰值信噪比(PSNR)改善了1~4dB,同时,该算法压缩率较高,运算速度较快,满足实时传输的需求。
A block-based compression method for computer screen image was proposed to realize real-time computer screen image transmission. Each block of image was classified into three categories: text/graphic block, hybrid block and picture block according to the different characters on color and texture. Text/ graphic block was compressed with a lossless coding algorithm integrating palette coding, Hextile coding and LZW algorithm. Picture block was compressed with lossy JPEG. Both lossless and lossy methods were used on hybrid block. Experimental results showed that the method exactly distinguished between the text/graphics and the picture information. The peak signal-to-noise ratio (PSNR) of the reconstructed image was improved 1~4 dB compared with other methods. The method achieved good compression ratio with low computation, which satisfied the real-time requirement.