A no-reference method for JPEG image compression impairment evaluation

Visual quality measurement of the compressed image is of importance to evaluate the effective of image compression algorithm. In this paper, we adopt a no-reference method to predict the perceived quality of JPEG compressed images. We claim that the restored information, which is got through postfiltering in shifted windows of image blocks, can replace the reference image to be compared with the distorted image in structure similarity index. The performance of the proposed measure is validated on a database of JPEG compressed images [1]. The experimental results prove that our method has a good performance on predicting the image quality compressed by JPEG standard.