A novel image quality evaluation method based on two-dimensional information entropy

Abstract

In image transmission, with the consideration of system parameters, such as bandwidth, transmission delay and power consumption, the image compression coding is often used to reduce the amount of transmitted data. The formation mechanism of block effect in H.264 coding is analyzed. Taking H.264 coding as an example, the spatial domain image enhancement processing is adopted in the paper to obtain the two-dimensional information entropy covering the gray correlation and image edge feature information in adjacent spatial domain. By referring to the two-dimensional information entropy evaluation value of original and coded image, the image quality evaluation result is consistent with PSNR method used in common situations. Compared with the traditional PSNR evaluation method, the two-dimensional information entropy evaluation method only refers to the two-dimensional information entropy evaluation value of the original image, and the amount of data required is very low. Therefor, it is more suitable for the engineering implementation of real-time image quality evaluation in image transmission system. At the same time, with the edge feature information of the image used in the two-dimensional information entropy, the entropy method is closer to the actual feeling of human eyes than PSNR method.