CIC: Circular Image Compression

2024-07-18

arXiv.org

AI Summary

Existing learned image compression (LIC) methods suffer significant performance degradation under out-of-distribution (e.g., out-of-sample or out-of-domain) testing. To address this, we propose a closed-loop circular image compression framework—the first to introduce nonlinear feedback control into LIC—dynamically calibrating latent representations via an encoder-decoder feedback loop to bridge the train-test distribution gap. We theoretically prove that the steady-state reconstruction error converges to zero. The method is plug-and-play, requiring no retraining and compatible with any serial compression architecture. Leveraging automatic control modeling and Taylor-series-based error analysis, our approach consistently outperforms five state-of-the-art open-source LIC methods across five public benchmarks. Notably, it delivers substantial reconstruction quality improvements on challenging cases—including dark backgrounds, high-contrast regions, sharp edges, and complex textures—demonstrating superior generalization and robustness.

Technology Category

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

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Learned image compression (LIC) is currently the cutting-edge method. However, the inherent difference between testing and training images of LIC results in performance degradation to some extent. Especially for out-of-sample, out-of-distribution, or out-of-domain testing images, the performance of LIC dramatically degraded. Classical LIC is a serial image compression (SIC) approach that utilizes an open-loop architecture with serial encoding and decoding units. Nevertheless, according to the theory of automatic control, a closed-loop architecture holds the potential to improve the dynamic and static performance of LIC. Therefore, a circular image compression (CIC) approach with closed-loop encoding and decoding elements is proposed to minimize the gap between testing and training images and upgrade the capability of LIC. The proposed CIC establishes a nonlinear loop equation and proves that steady-state error between reconstructed and original images is close to zero by Taylor series expansion. The proposed CIC method possesses the property of Post-Training and plug-and-play which can be built on any existing advanced SIC methods. Experimental results on five public image compression datasets demonstrate that the proposed CIC outperforms five competing state-of-the-art open-source SIC algorithms in reconstruction capacity. Experimental results further show that the proposed method is suitable for out-of-sample testing images with dark backgrounds, sharp edges, high contrast, grid shapes, or complex patterns.

Problem

• Improves learned image compression for out-of-distribution images
• Proposes closed-loop architecture to reduce training-testing gap
• Enhances reconstruction on images with complex patterns and backgrounds

Innovation

• Closed-loop architecture for image compression
• Minimizes gap between testing and training images
• Plug-and-play upgrade for existing compression methods