Low Complexity Image Quality Assessment Using Frequency Domain Transforms

This paper proposes a low complexity image quality assessment method based on frequency domain transforms. The proposed scheme is easy to implement and reliably estimates image quality across various distortions types including images compressed using frequency domain transforms (JPEG and JPEG2000). The performance of the two proposed metrics (Q-DCT and Q-DWT) is compared with two full-reference metrics, the PSNR and another recent state-of-the-art measure. The proposed metrics provide a global numerical image quality measure but can also be used to obtain local image quality statistics. Preliminary results indicate that the proposed scheme can be adapted to perform well in reduced reference frameworks making it well-suited for real-time applications.