Learning Adaptive Lighting via Channel-Aware Guidance

2024-12-02

arXiv.org

AI Summary

Existing illumination-related vision tasks—such as HDR reconstruction and exposure correction—are typically modeled in isolation, overlooking intrinsic illumination characteristics differing across RGB channels in both spatial and frequency domains. To address this, we propose LALNet, the first unified framework incorporating dual-domain channel modulation and a light-guided attention (LGA) mechanism. LALNet jointly models channel-specific illumination variations and cross-channel visual consistency: it employs wavelet transforms for frequency-domain disentanglement, integrates a visual state-space model to enhance long-range dependency modeling, and adopts color-separation–color-mixing feature fusion to preserve chromatic fidelity. Evaluated on four mainstream illumination tasks, LALNet achieves new state-of-the-art performance, improving average PSNR by 1.2–2.8 dB, reducing parameter count by 37%, and accelerating inference by 2.1×. An online demonstration system is publicly released.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Learning lighting adaption is a key step in obtaining a good visual perception and supporting downstream vision tasks. There are multiple light-related tasks (e.g., image retouching and exposure correction) and previous studies have mainly investigated these tasks individually. However, we observe that the light-related tasks share fundamental properties: i) different color channels have different light properties, and ii) the channel differences reflected in the time and frequency domains are different. Based on the common light property guidance, we propose a Learning Adaptive Lighting Network (LALNet), a unified framework capable of processing different light-related tasks. Specifically, we introduce the color-separated features that emphasize the light difference of different color channels and combine them with the traditional color-mixed features by Light Guided Attention (LGA). The LGA utilizes color-separated features to guide color-mixed features focusing on channel differences and ensuring visual consistency across channels. We introduce dual domain channel modulation to generate color-separated features and a wavelet followed by a vision state space module to generate color-mixed features. Extensive experiments on four representative light-related tasks demonstrate that LALNet significantly outperforms state-of-the-art methods on benchmark tests and requires fewer computational resources. We provide an anonymous online demo at https://xxxxxx2025.github.io/LALNet/.

Problem

• Addresses multiple light-related tasks efficiently
• Explores channel differences in spatial and frequency domains
• Improves visual consistency across color channels

Innovation

• Channel-aware Learning Adaptive Lighting Network (LALNet)
• Light Guided Attention (LGA) for channel differences
• Dual domain channel modulation for color features