Window-based Channel Attention for Wavelet-enhanced Learned Image Compression

2024-09-21

Asian Conference on Computer Vision

AI Summary

To address the insufficient global modeling capability and poor detail preservation in learned image compression, this paper proposes a spatial-channel hybrid attention framework. First, it introduces window partitioning into channel attention, designing Windowed Channel Attention (WCA) to enhance long-range inter-channel dependency modeling. Second, it integrates a joint spatial-channel attention mechanism with Discrete Wavelet Transform (DWT)-based frequency-aware downsampling to expand the effective receptive field. Within a Transformer-based architecture, the method simultaneously captures large-receptive-field global correlations and small-receptive-field local detail fidelity. Evaluated on four standard benchmark datasets, the proposed approach achieves an average BD-rate reduction of 22.39% over VTM-23.1, establishing new state-of-the-art performance in learned image compression.

Technology Category

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

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Learned Image Compression (LIC) models have achieved superior rate-distortion performance than traditional codecs. Existing LIC models use CNN, Transformer, or Mixed CNN-Transformer as basic blocks. However, limited by the shifted window attention, Swin-Transformer-based LIC exhibits a restricted growth of receptive fields, affecting the ability to model large objects for image compression. To address this issue and improve the performance, we incorporate window partition into channel attention for the first time to obtain large receptive fields and capture more global information. Since channel attention hinders local information learning, it is important to extend existing attention mechanisms in Transformer codecs to the space-channel attention to establish multiple receptive fields, being able to capture global correlations with large receptive fields while maintaining detailed characterization of local correlations with small receptive fields. We also incorporate the discrete wavelet transform into our Spatial-Channel Hybrid (SCH) framework for efficient frequency-dependent down-sampling and further enlarging receptive fields. Experiment results demonstrate that our method achieves state-of-the-art performances, reducing BD-rate by 18.54%, 23.98%, 22.33%, and 24.71% on four standard datasets compared to VTM-23.1.

Problem

• Image Compression
• Visual Information Processing
• Detail Retention

Innovation

• Window-based Channel Attention
• Discrete Wavelet Transform
• Swin-Transformer Enhancement