A perceptual no-reference blockiness metric for JPEG images

In this paper we present a novel no-reference (NR) metric to measure block impairment (blockiness) in JPEG-coded images. The proposed metric integrates several key human visual sensitivity factors such as edge amplitude, edge length, background activity and background luminance to evaluate the effect of block edge impairment on perceived image quality. The subjective test results of our metric is compared with the Wang-Bovik's NR blockiness metric. The results show that the proposed metric correlates well with the mean opinion score (MOS) than Wang-Bovik's blockiness metric. Further this metric can be extended to predict the quality of the MPEG/H.26x compressed videos.