High Dynamic Range Video Compression: A Large-Scale Benchmark Dataset and A Learned Bit-depth Scalable Compression Algorithm

2025-03-01

AI Summary

HDR video learning-based compression is hindered by the scarcity of large-scale, high-quality training data. To address this, we introduce HDRVD2K—the first open-source, large-scale HDR video benchmark—and propose LBSVC, the first end-to-end learned video compression framework supporting bit-depth scalability. Its core innovation is the Bit-depth Enhancement Module (BEM), a compression-friendly component that explicitly models and exploits bit-depth redundancy across multi-dynamic-range videos, moving beyond conventional spatio-temporal prediction paradigms. By incorporating dynamic-range priors, BEM jointly optimizes HDR reconstruction fidelity and coding efficiency. Experiments demonstrate that LBSVC achieves a 1.8 dB PSNR gain and over 35% bitrate reduction on HDR video compared to state-of-the-art methods. Both the code and the HDRVD2K dataset are publicly released.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Learning on the Edge & Model Compression
• Data Mining & Knowledge Management: Data Compression

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Recently, learned video compression (LVC) is undergoing a period of rapid development. However, due to absence of large and high-quality high dynamic range (HDR) video training data, LVC on HDR video is still unexplored. In this paper, we are the first to collect a large-scale HDR video benchmark dataset, named HDRVD2K, featuring huge quantity, diverse scenes and multiple motion types. HDRVD2K fills gaps of video training data and facilitate the development of LVC on HDR videos. Based on HDRVD2K, we further propose the first learned bit-depth scalable video compression (LBSVC) network for HDR videos by effectively exploiting bit-depth redundancy between videos of multiple dynamic ranges. To achieve this, we first propose a compression-friendly bit-depth enhancement module (BEM) to effectively predict original HDR videos based on compressed tone-mapped low dynamic range (LDR) videos and dynamic range prior, instead of reducing redundancy only through spatio-temporal predictions. Our method greatly improves the reconstruction quality and compression performance on HDR videos. Extensive experiments demonstrate the effectiveness of HDRVD2K on learned HDR video compression and great compression performance of our proposed LBSVC network. Code and dataset will be released in https://github.com/sdkinda/HDR-Learned-Video-Coding.

Problem

• Lack of large-scale HDR video training data.
• Need for efficient HDR video compression methods.
• Improving reconstruction quality and compression performance.

Innovation

• Developed large-scale HDR video dataset HDRVD2K
• Proposed learned bit-depth scalable compression network
• Introduced bit-depth enhancement module for HDR reconstruction