A Multi-Scale Spatial Attention-Based Zero-Shot Learning Framework for Low-Light Image Enhancement

2025-06-23

AI Summary

Addressing low-light image enhancement under unpaired training data, this paper proposes LucentVisionNet—a zero-shot enhancement framework. Methodologically, it integrates a multi-scale spatial attention mechanism with a deep curve estimation network to enable fine-grained joint modeling of brightness and structural details. It innovatively introduces a no-reference perceptual quality loss grounded in human visual system characteristics and designs a composite six-term loss function coupled with a cyclic enhancement strategy, significantly improving generalization and semantic fidelity. Extensive experiments on multiple benchmark datasets demonstrate that LucentVisionNet consistently outperforms state-of-the-art supervised, unsupervised, and zero-shot methods, achieving new SOTA results in PSNR, SSIM, and user study metrics. Moreover, it maintains high computational efficiency, making it suitable for real-world applications such as mobile photography, intelligent surveillance, and autonomous driving.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Security and Privacy: Privacy-enhancing technologies

Abstract

Low-light image enhancement remains a challenging task, particularly in the absence of paired training data. In this study, we present LucentVisionNet, a novel zero-shot learning framework that addresses the limitations of traditional and deep learning-based enhancement methods. The proposed approach integrates multi-scale spatial attention with a deep curve estimation network, enabling fine-grained enhancement while preserving semantic and perceptual fidelity. To further improve generalization, we adopt a recurrent enhancement strategy and optimize the model using a composite loss function comprising six tailored components, including a novel no-reference image quality loss inspired by human visual perception. Extensive experiments on both paired and unpaired benchmark datasets demonstrate that LucentVisionNet consistently outperforms state-of-the-art supervised, unsupervised, and zero-shot methods across multiple full-reference and no-reference image quality metrics. Our framework achieves high visual quality, structural consistency, and computational efficiency, making it well-suited for deployment in real-world applications such as mobile photography, surveillance, and autonomous navigation.

Problem

• Enhancing low-light images without paired training data
• Overcoming limitations of traditional and deep learning methods
• Improving generalization with multi-scale attention and recurrent strategy

Innovation

• Multi-scale spatial attention for fine enhancement
• Deep curve estimation network integration
• Recurrent enhancement with composite loss function