Residual Learning and Filtering Networks for End-to-End Lossless Video Compression

2025-03-11

AI Summary

Existing learning-based video compression methods suffer from inaccurate motion estimation and weak motion compensation structures, resulting in large reconstruction errors and suboptimal rate-distortion (RD) performance. To address these limitations, this paper proposes an end-to-end learnable lossless video compression framework. First, we introduce a novel residual skip-connected autoencoder to efficiently compress motion information. Second, we jointly optimize the motion vector prediction network and the residual frame filtering network. Third, we propose a reference frame fine-tuning mechanism with a learnable buffer, coupled with a PReLU-enhanced motion compensation structure. The entire framework is trained via end-to-end RD-optimized joint learning. Experimental results demonstrate significant improvements over state-of-the-art methods on standard benchmarks—including HEVC (Classes B/C/D), UVG, VTL, and MCL-JCV—achieving new SOTA performance in both reconstruction quality and RD efficiency.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Existing learning-based video compression methods still face challenges related to inaccurate motion estimates and inadequate motion compensation structures. These issues result in compression errors and a suboptimal rate-distortion trade-off. To address these challenges, this work presents an end-to-end video compression method that incorporates several key operations. Specifically, we propose an autoencoder-type network with a residual skip connection to efficiently compress motion information. Additionally, we design motion vector and residual frame filtering networks to mitigate compression errors in the video compression system. To improve the effectiveness of the motion compensation network, we utilize powerful nonlinear transforms, such as the Parametric Rectified Linear Unit (PReLU), to delve deeper into the motion compensation architecture. Furthermore, a buffer is introduced to fine-tune the previous reference frames, thereby enhancing the reconstructed frame quality. These modules are combined with a carefully designed loss function that assesses the trade-off and enhances the overall video quality of the decoded output. Experimental results showcase the competitive performance of our method on various datasets, including HEVC (sequences B, C, and D), UVG, VTL, and MCL-JCV. The proposed approach tackles the challenges of accurate motion estimation and motion compensation in video compression, and the results highlight its competitive performance compared to existing methods.

Problem

• Addresses inaccurate motion estimates in video compression.
• Improves motion compensation structures to reduce compression errors.
• Enhances rate-distortion trade-off for better video quality.

Innovation

• Autoencoder network with residual skip connection
• Motion vector and residual frame filtering networks
• Nonlinear transforms like PReLU for motion compensation