This paper proposed a metric for video quality evaluation based on the HVS properties. In different temporal region, people would have different sensibility to the distortion of video. So the metric modified the traditional edge detection method (Yousei's model which was one of the recommendations in ITU-T J.247) by taking the temporal masking effect into account. Firstly, the temporal gradient had been done to find out the temporal distribution. Then, the weight had been calculated based on a HVS model to emulate the temporal property. Finally, these weights were used to modify the spatial value. By doing these, it would be able to combine the spatial and temporal detection together and proposed a new metric for video quality measurement.
展开▼
