Bridging Robustness and Efficiency: Real-Time Low-Light Enhancement via Attention U-Net GAN

TL;DR

This paper introduces a hybrid Attention U-Net GAN for low-light image enhancement that balances high perceptual quality with real-time inference speed, outperforming existing models in both texture recovery and efficiency.

Contribution

The paper presents a novel hybrid Attention U-Net GAN that achieves high-fidelity texture recovery in low-light images with near real-time speed, bridging the gap between diffusion models and CNN-based methods.

Findings

Achieves LPIPS score of 0.112 on SID dataset.

Inference latency of 0.06 seconds per image.

40x speedup over latent diffusion models.

Abstract

Recent advancements in Low-Light Image Enhancement (LLIE) have focused heavily on Diffusion Probabilistic Models, which achieve high perceptual quality but suffer from significant computational latency (often exceeding 2-4 seconds per image). Conversely, traditional CNN-based baselines offer real-time inference but struggle with "over-smoothing," failing to recover fine structural details in extreme low-light conditions. This creates a practical gap in the literature: the lack of a model that provides generative-level texture recovery at edge-deployable speeds. In this paper, we address this trade-off by proposing a hybrid Attention U-Net GAN. We demonstrate that the heavy iterative sampling of diffusion models is not strictly necessary for texture recovery. Instead, by integrating Attention Gates into a lightweight U-Net backbone and training within a conditional adversarial framework, we can approximate the high-frequency fidelity of generative models in a single forward pass. Extensive experiments on the SID dataset show that our method achieves a best-in-class LPIPS score of 0.112 among efficient models, significantly outperforming efficient baselines (SID, EnlightenGAN) while maintaining an inference latency of 0.06s. This represents a 40x speedup over latent diffusion models, making our approach suitable for near real-time applications.