Abstract:To reduce the redundancy among different color channels, e.g., YUV, previous methods usually adopt a linear model that tends to be oversimple for complex image content. We propose a neural-network-based method for cross-channel prediction in intra frame coding. The proposed network utilizes twofold cues, i.e., the neighboring reconstructed samples with all channels, and the co-located reconstructed samples with partial channels. Specifically, for YUV video coding, the neighboring samples with YUV are processed by several fully connected layers; the co-located samples with Y are processed by convolutional layers; and the proposed network fuses the twofold cues. We observe that the integration of twofold information is crucial to the performance of intra prediction of the chroma components. We have designed the network architecture to achieve a good balance between compression performance and computational efficiency. Moreover, we propose a transform domain loss for the training of the network. The transform domain loss helps obtain more compact representations of residues in the transform domain, leading to higher compression efficiency. The proposed method is plugged into HEVC and VVC test models to evaluate its effectiveness. Experimental results show that our method provides more accurate cross-channel intra prediction compared with previous methods. On top of HEVC, our method achieves on average 1.3%, 5.4%, and 3.8% BD-rate reductions for Y, Cb, and Cr on common test sequences, and on average 3.8%, 11.3%, and 9.0% BD-rate reductions for Y, Cb, and Cr on ultra-high-definition test sequences. On top of VVC, our method achieves on average 0.5%, 1.7%, and 1.3% BD-rate reductions for Y, Cb, and Cr on common test sequences.

Deep Intra Prediction by Jointly Exploiting Local and Non-Local Similarities

Intra Prediction Using Fully Connected Network for Video Coding

Coarse-to-fine Prediction with Local and Nonlocal Correlations for Intra Coding

Current Frame Priors Assisted Neural Network For Intra Prediction

Enhanced Intra Prediction with Recurrent Neural Network in Video Coding

Fully Connected Network-Based Intra Prediction for Image Coding.

Joint Local and Nonlocal Progressive Prediction for Versatile Video Coding

Progressive Spatial Recurrent Neural Network for Intra Prediction

Deep Inter Prediction Via Pixel-Wise Motion Oriented Reference Generation

Hybrid Intraprediction Based on Local and Nonlocal Correlations

Optimized Spatial Recurrent Network for Intra Prediction in Video Coding

Towards Lightweight Neural Network-based Chroma Intra Prediction for Video Coding

Chroma Intra Prediction with Lightweight Attention-Based Neural Networks

A new intra prediction method based on consistent luminance changes.

Spatial Information Refinement for Chroma Intra Prediction in Video Coding

Intra Prediction Using Multiple Reference Lines for Video Coding

Intra Prediction Based on Geometry Padding for Omnidirectional Video Coding

Convolutional Neural Networks Based Intra Prediction for HEVC

Neural Network-Based Enhancement to Inter Prediction for Video Coding

A Hybrid Neural Network for Chroma Intra Prediction

Neural-Network-Based Cross-Channel Intra Prediction