基于模糊推理的彩色图像帧内误码掩盖方法

本文研究了RGB彩色图像基于模糊推理的误码掩盖（FRBEC）方法。首先将RGB图像转换到YCbCr彩色空间,对Y分量图像的丢失块进行周围像素匹配以获得最佳匹配块,并把最佳匹配块的位置直接作为R、G、B分量图像对应的最佳匹配块位置,以减少计算量。然后分别对R、G、B3幅分量图像采用模糊推理对匹配的结果进行模糊分类,并对匹配准确度低的块进行改进的像素逐点修正。实验结果表明,FRBEC方法无论对于彩色和灰度的图像,都可以获得较高的峰值信噪比（PSNR）,不过彩色图像的PSNR一般比灰度图像要低一些。

Fuzzy Reasoning Based Intra Frame Error Concealment For Block Based Color Images

A new technique of the fuzzy reasoning based intra frame error concealment （FRBEC） for the block-based RGB color images was proposed in this paper. First, RGB image was converted to YCbCr color space. The pixel matching was applied to the Y component image to get the location of the optimal matching block. In order to reduce computation, we considered this location as the corresponding location of the optimal matching block of R, G, B component image. Then the fuzzy reasoning was applied to R, G, B component image respectively to fuzzy classify the results of pixel matching. And then the developed Pixel correction was processed to modify the pixels of the blocks with low recover exactness degree （RED）. The experiment results show that the FRBEC scheme for both color images and gray images gives good results. The PSNR of the color images is lower than that of the gray images.