Blind quality assessment of videos using a model of natural scene statistics and motion coherency
Michele A. Saad, Alan Conrad Bovik · 2012
We propose a no-reference algorithm for video quality evaluation. The algorithm relies on a natural scene statistics (NSS) model of video DCT coefficients as well as a temporal model of motion coherency. The proposed framework is tested on the LIVE VQA database, and shown to correlate well with human visual judgments of quality.