Modeling the hybrid temporal and spatial resolutions effect for web video quality evaluation

Understanding and modelling the users´ perceptual quality of a video are the key steps towards improving multimedia service provision. In this paper, we take an analytical approach to study the joint impact of spatial and temporal resolution on the perceptual video quality. First, we map the subjective quality data as an evaluation function in terms of the spatial resolution and frame rate, which reflect the major features of web video quality. Second, we use ε-Support Vector Regression (ε-SVR) with Radial Basis Function (RBF) kernel to give a more accurate hybrid temporal-spatial quality evaluation model so as to predict the perceptual video quality. The comprehensive subjective quality experiments were carried out to construct and validate this model. The experimental results demonstrated the effectiveness of the proposed model. Besides, the quality model can be easily deployed in a practical multimedia delivery system.