Hybrid bitstream-based video quality assessment method for scalable video coding

摘要

We propose a quality assessment method from decoding parameters of compressed bitstreams by scalable video coding (SVC) as a hybrid/bitstream category. Conventional video quality assessment methods evaluate the video quality of degraded videos after full reconstruction. However, the proposed assessment method quantifies video quality of SVC not only with reconstructed videos for the base layer but also with decoding parameters for the enhancement layer. The proposed algorithm assesses the enhanced quality degree of the enhancement layers with statistics of various coding parameters for the enhancement layers with respect to the quality of the base layer. The accuracy of the proposed algorithm is 23% higher than those of conventional algorithms in terms of Pearson correlation. Furthermore, the proposed algorithm has significantly lower computational complexity than conventional methods.