Title: New spatiotemporal method for assessing video quality

Authors: Woei-Tan Loh; David Boon Liang Bong

Addresses: Faculty of Engineering, Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia ' Faculty of Engineering, Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia

Abstract: The existence of temporal effects and temporal distortions in a video differentiate the way it is assessed from an image. Temporal effects and distortions can enhance or depress the visibility of spatial effects in a video. Thus, the temporal part of videos plays a significant role in determining the video quality. In this study, a spatiotemporal video quality assessment (VQA) method is proposed due to the importance of temporal effects and distortions in assessing video quality. Instead of measuring the frame quality on a frame basis, the quality of several averaged frames is measured. The proposed spatiotemporal VQA method is significantly improved compared with image quality assessment (IQA) methods applied on a frame basis. When combined with IQA methods, the proposed spatiotemporal VQA method has comparable performance with state-of-the-art VQA methods. The computational complexity of the proposed temporal method is also lower when compared with current VQA methods.

Keywords: video; frames; video quality; spatial effects; temporal effects; temporal distortions; spatiotemporal; average; video quality assessment; VQA; image quality assessment; IQA; computational complexity.

DOI: 10.1504/IJCVR.2018.093077

International Journal of Computational Vision and Robotics, 2018 Vol.8 No.3, pp.336 - 348

Accepted: 15 Nov 2017
Published online: 07 Jul 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article