SIMI: Self-information Mining Network for Low-light Image Enhancement

2026-05-08

AI Summary

This work addresses the challenge of low-light image enhancement, where poor signal-to-noise ratio severely degrades downstream visual tasks. The authors propose an unsupervised self-information mining network that, for the first time, integrates bit-plane decomposition into an unsupervised low-light enhancement framework. By decomposing an image into multiple bit-plane components, the method effectively uncovers intrinsic structural and textural information without requiring paired training data. Extensive experiments demonstrate that the proposed approach achieves state-of-the-art performance across multiple standard benchmarks, while simultaneously offering significantly faster convergence and reduced computational overhead, thereby demonstrating both efficacy and practicality.

Abstract

Poor lighting conditions significantly impact image quality, posing substantial challenges for image editing and visualization. Many existing enhancement methods aim at proposing complex models while neglecting the intrinsic information contained within low-light images. In this work, we propose the Self-Information Mining (SIMI) network, an innovative unsupervised framework that decomposes low-light images into multiple components based on bit-plane decomposition. Our approach allows mining intrinsic information without relying on external data. This not only accelerates model convergence but also improves performance and reduces computational overhead. The unsupervised nature of our method facilitates real-world applicability. Experiments conducted on standard benchmarks demonstrate that SIMI achieves state-of-the-art performance.

Problem

• low-light image enhancement
• image quality
• intrinsic information
• poor lighting conditions
• unsupervised learning

Innovation

• Self-Information Mining
• Bit-plane Decomposition
• Unsupervised Learning
• Low-light Image Enhancement
• Intrinsic Information