A novel reduced reference image quality analysis metric for JPEG compressed images based on image segmentation

In this article, we propose a reduced reference image quality analysis (IQA) metric that requires very little amount of information from the reference image for calculating the quality of the corresponding distorted image. The proposed metric determines the quality of an image distorted by the JPEG image compression scheme by calculating the number of connected components, where a connected component is a cluster of pixels with same color. It can be observed that by increasing the compression ratio, the JPEG scheme reduces the number of connected components in the image. In this research, the quality of a distorted image is derived from the difference between the number of connected components in the reference image and the distorted image. The fewer connected components a distorted image has, the lower the quality of the image is. Comparison between the proposed metric and the existing reduced reference IQA methods reveals that the proposed metric can efficiently reduce the information represented by the reference image with reasonable accuracy.