Towards Lightest Low-Light Image Enhancement Architecture for Mobile Devices

2025-07-06

AI Summary

Mobile low-light image enhancement faces challenges in balancing model lightweighting and real-time inference while heavily relying on large-scale annotated data. Method: This paper proposes LiteIE, an ultra-lightweight unsupervised framework featuring: (i) a backbone-agnostic feature extractor with only two convolutional layers; (ii) a parameter-free, parameter-sharing iterative restoration module; and (iii) an unsupervised loss function integrating exposure control, edge-aware smoothing, and multi-scale color consistency. Contribution/Results: LiteIE achieves high-quality enhancement with only 58 parameters—the first such result under extreme structural simplification—attaining 19.04 dB PSNR on the LOL dataset, surpassing state-of-the-art (SOTA) methods by 1.4 dB and reducing parameter count to just 0.07% of SOTA. On Snapdragon 8 Gen 3, it enables real-time 4K image processing at 30 FPS, significantly improving cross-scene generalization and deployment efficiency.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Privacy-enhancing technologies

Abstract

Real-time low-light image enhancement on mobile and embedded devices requires models that balance visual quality and computational efficiency. Existing deep learning methods often rely on large networks and labeled datasets, limiting their deployment on resource-constrained platforms. In this paper, we propose LiteIE, an ultra-lightweight unsupervised enhancement framework that eliminates dependence on large-scale supervision and generalizes well across diverse conditions. We design a backbone-agnostic feature extractor with only two convolutional layers to produce compact image features enhancement tensors. In addition, we develop a parameter-free Iterative Restoration Module, which reuses the extracted features to progressively recover fine details lost in earlier enhancement steps, without introducing any additional learnable parameters. We further propose an unsupervised training objective that integrates exposure control, edge-aware smoothness, and multi-scale color consistency losses. Experiments on the LOL dataset, LiteIE achieves 19.04 dB PSNR, surpassing SOTA by 1.4 dB while using only 0.07% of its parameters. On a Snapdragon 8 Gen 3 mobile processor, LiteIE runs at 30 FPS for 4K images with just 58 parameters, enabling real-time deployment on edge devices. These results establish LiteIE as an efficient and practical solution for low-light enhancement on resource-limited platforms.

Problem

• Balancing visual quality and computational efficiency for mobile low-light enhancement
• Eliminating dependence on large networks and labeled datasets
• Enabling real-time 4K image enhancement on resource-limited devices

Innovation

• Ultra-lightweight unsupervised enhancement framework
• Backbone-agnostic feature extractor with two layers
• Parameter-free Iterative Restoration Module