ERIENet: An Efficient RAW Image Enhancement Network under Low-Light Environment

2025-12-17

AI Summary

Existing RAW-domain low-light enhancement methods suffer from model bloat, slow inference, and ineffective exploitation of the green-channel prior. To address these issues, we propose a lightweight fully parallel multi-scale architecture that eliminates sequential multi-scale processing to significantly reduce computational redundancy. We introduce a channel-aware residual dense block and a novel lightweight green-channel guidance branch, explicitly modeling the green channel’s superior signal-to-noise ratio and structural fidelity. Our method further incorporates end-to-end RAW-domain supervision and efficient convolutional modules. Extensive experiments demonstrate state-of-the-art performance on benchmark datasets including SID and LOL. On an RTX 3090 GPU, our model achieves over 146 FPS inference speed for 4K images—striking an unprecedented balance between reconstruction quality and real-time efficiency.

Technology Category

• Machine Learning: Efficient ML / Green AI
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Security and Privacy: Privacy-enhancing technologies

Abstract

RAW images have shown superior performance than sRGB images in many image processing tasks, especially for low-light image enhancement. However, most existing methods for RAW-based low-light enhancement usually sequentially process multi-scale information, which makes it difficult to achieve lightweight models and high processing speeds. Besides, they usually ignore the green channel superiority of RAW images, and fail to achieve better reconstruction performance with good use of green channel information. In this work, we propose an efficient RAW Image Enhancement Network (ERIENet), which parallelly processes multi-scale information with efficient convolution modules, and takes advantage of rich information in green channels to guide the reconstruction of images. Firstly, we introduce an efficient multi-scale fully-parallel architecture with a novel channel-aware residual dense block to extract feature maps, which reduces computational costs and achieves real-time processing speed. Secondly, we introduce a green channel guidance branch to exploit the rich information within the green channels of the input RAW image. It increases the quality of reconstruction results with few parameters and computations. Experiments on commonly used low-light image enhancement datasets show that ERIENet outperforms state-of-the-art methods in enhancing low-light RAW images with higher effiency. It also achieves an optimal speed of over 146 frame-per-second (FPS) for 4K-resolution images on a single NVIDIA GeForce RTX 3090 with 24G memory.

Problem

• Enhances low-light RAW images efficiently with real-time processing.
• Utilizes green channel information to improve image reconstruction quality.
• Addresses computational cost and speed limitations in existing methods.

Innovation

• Parallel multi-scale processing with efficient convolution modules
• Green channel guidance branch for enhanced reconstruction
• Real-time processing speed with low computational cost