This is the source code of Video Quality Assessment Based on Human Visual Perception.
It mainly contains the network structure and training files. The overall structure of our method is described in CNN3D.py.
More training details and our pretrained models will be released as soon as possible.