An Efficient Low-Complexity Convolutional Neural Network Filter

Convolutional neural network (CNN) filters have achieved significant performance in video artifacts reduction. However, the high complexity of existing methods makes them difficult to be applied in actual usage. In this paper, an efficient low-complexity CNN filter is proposed. We utilized depth separable convolution merged with the batch normalization as the backbone of our proposed CNN filter and presented a frame-level residual mapping (RM) to use one network to filter both Intra and Inter samples. It is known that there will be an over smoothing problem for the Inter frames if we directly use the filter trained with Intra samples. In this paper, the proposed RM can effectively solve the over smoothing problem. Besides, RM is flexible and can be combined with other learning-based filters. The experimental results show that our proposed method achieves a significant BD-rate reduction than H.265/HEVC. The experiments show the proposed network achieves about 1.2% BD-rate reduction and 79.1% decrease in FLOPs than VR-CNN. Our performance is better with less complexity than the previous work. The measurement on H.266/VVC and ablation studies also ensure the effectiveness of the proposed method.