Fast VVC Intra Coding Algorithm Based on Luminance-Chrominance Feature and Deep Learning
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
To address the issues of high intra coding complexity of the new-generation video coding standard Versatile Video Coding (VVC) and the neglect of chroma components' value in existing research, this paper proposes a deep learning-based video coding optimization algorithm utilizing luma and chroma features. The algorithm aims to reduce the intra prediction mode decision overhead of VVC while improving prediction accuracy. Firstly, we constructs a large-scale structured Coding Unit (CU)-level dataset. Secondly, we designs a multi-channel feature fusion strategy, builds a model based on neural networks, and comparatively analyzes the performance improvement of intra mode prediction brought by the joint representation of luma (Y), blue-difference (U), and red-difference (V) components. Experimental results show that after introducing chrominance components, the model's intra-frame mode prediction accuracy on Class A and Class B video sequences is improved by approximately 5%. This algorithm can balance the complexity and performance of VVC coding, providing technical support for efficient coding in scenarios such as ultra-high definition video and real-time video communication.
Fields of Study & Subjects
Video Coding and Compression Technologies
Advanced Data Compression Techniques
Image and Video Quality Assessment
Publication details
- Journal / source
- International Conference on Communication Networks and Machine Learning (CNML)
- Pages
- 460–464
- Publication date
- 30 January 2026
- Language
- EN