3DM-WeConvene: Learned Image Compression with 3D Multi-Level Wavelet-Domain Convolution and Entropy Model

2025-04-07

AI Summary

To address insufficient modeling of frequency-domain correlations and suboptimal rate-distortion (RD) performance in learning-based image compression, this paper proposes the first end-to-end framework integrating 3D multi-level wavelet transforms. Its core contributions are: (1) a 3D wavelet-domain convolutional layer (3DM-WeConv) that jointly suppresses spatial, channel-wise, and frequency-domain redundancies; and (2) a wavelet-domain channel-wise autoregressive entropy model (3DWeChARM), enabling low-frequency prior-guided high-frequency coding and two-stage rate imbalance optimization. The framework incorporates both 5/3 and 9/7 wavelet transforms, subband-adaptive convolutions, inverse 3D discrete wavelet transform (DWT), and slice-based entropy coding, trained via a two-stage weighted strategy. Experimental results on Kodak, Tecnick100, and CLIC datasets show BD-rate reductions of 12.24%, 15.51%, and 12.97%, respectively, over H.266/VVC—demonstrating substantial improvements in RD performance and computational efficiency for high-resolution image compression.

Technology Category

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

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Learned image compression (LIC) has recently made significant progress, surpassing traditional methods. However, most LIC approaches operate mainly in the spatial domain and lack mechanisms for reducing frequency-domain correlations. To address this, we propose a novel framework that integrates low-complexity 3D multi-level Discrete Wavelet Transform (DWT) into convolutional layers and entropy coding, reducing both spatial and channel correlations to improve frequency selectivity and rate-distortion (R-D) performance. Our proposed 3D multi-level wavelet-domain convolution (3DM-WeConv) layer first applies 3D multi-level DWT (e.g., 5/3 and 9/7 wavelets from JPEG 2000) to transform data into the wavelet domain. Then, different-sized convolutions are applied to different frequency subbands, followed by inverse 3D DWT to restore the spatial domain. The 3DM-WeConv layer can be flexibly used within existing CNN-based LIC models. We also introduce a 3D wavelet-domain channel-wise autoregressive entropy model (3DWeChARM), which performs slice-based entropy coding in the 3D DWT domain. Low-frequency (LF) slices are encoded first to provide priors for high-frequency (HF) slices. A two-step training strategy is adopted: first balancing LF and HF rates, then fine-tuning with separate weights. Extensive experiments demonstrate that our framework consistently outperforms state-of-the-art CNN-based LIC methods in R-D performance and computational complexity, with larger gains for high-resolution images. On the Kodak, Tecnick 100, and CLIC test sets, our method achieves BD-Rate reductions of -12.24%, -15.51%, and -12.97%, respectively, compared to H.266/VVC.

Problem

• Reducing frequency-domain correlations in learned image compression
• Improving rate-distortion performance with 3D wavelet transforms
• Enhancing entropy coding efficiency in wavelet domains

Innovation

• 3D multi-level wavelet-domain convolution layer
• 3D wavelet-domain channel-wise autoregressive entropy model
• Two-step training strategy for balance