【摘要】 提出一种专门针对彩色人头图像压缩编码的新方法。该方法利用彩色人头图像在各个特征区域三个色彩分量之间的色彩相关性,建立其中一个色彩分量与另两个色彩分量之间的相关模板,并只需对第一个色彩分量进行量化编码,解码时重构这个色彩分量后,加上极少量的相关模板系数就可解出另两个色彩分量,从而恢复图像。实验结果表明,该方法不仅可以大大提高编码效率和压缩比,并且能获得较满意的重构图像效果。更多还原

【Abstract】 This paper presents a new method specially for color head image compression coding. This method establishes the correlation template between one color component and the other two color components by using color correlation between three color components in each feature regions of a color head image. And it’s only necessary to quantize and code the first color component. During decoding, this color component is reconstructed. Adding a few correlation template coefficients, the other two color components can be calculated out to reconstruct the image. The experimental results show this method not only improve the coding efficient and compression ratio but also obtain satisfactory effect of the reconstructed image.更多还原