SurroundNet: Towards Effective Low-Light Image Enhancement

TL;DR

SurroundNet tackles the low-light enhancement problem with a lightweight CNN that embeds an Adaptive Surround Function to learn illumination estimation and adaptive Retinex Blocks to operate in feature space. By combining a Low-Exposure Denoiser and efficient channel attention, it achieves competitive state-of-the-art results with under 150K parameters. The approach is validated on real and synthetic datasets, showing strong PSNR/SSIM and perceptual metrics while maintaining a small model footprint. The work highlights the value of principled illumination priors and frequency-aware processing for robust, efficient low-light enhancement in practical applications.

Abstract

Although Convolution Neural Networks (CNNs) has made substantial progress in the low-light image enhancement task, one critical problem of CNNs is the paradox of model complexity and performance. This paper presents a novel SurroundNet which only involves less than 150$K$ parameters (about 80-98 percent size reduction compared to SOTAs) and achieves very competitive performance. The proposed network comprises several Adaptive Retinex Blocks (ARBlock), which can be viewed as a novel extension of Single Scale Retinex in feature space. The core of our ARBlock is an efficient illumination estimation function called Adaptive Surround Function (ASF). It can be regarded as a general form of surround functions and be implemented by convolution layers. In addition, we also introduce a Low-Exposure Denoiser (LED) to smooth the low-light image before the enhancement. We evaluate the proposed method on the real-world low-light dataset. Experimental results demonstrate that the superiority of our submitted SurroundNet in both performance and network parameters against State-of-the-Art low-light image enhancement methods. Code is available at https: github.com/ouc-ocean-group/SurroundNet.

Figures (12)

• Figure 1: Visual comparisons with two typical low-light image enhancement methods. The right side shows the zoom-in images of the selected small square area. (Image from LOL dataset). The proposed Surround achieves remarkable enhancement in brightness, color, and sharpness, while the other two methods generate image artifacts, otherwise low color saturation result.
• Figure 2: The top-left sub-figure is the comparison of three kind surround functions: inverse square $1/r^2$, exponential $e^{-|r|/c}$, and Gaussian $e^{-(r/\sigma)^2}$, where $c=30$, $\sigma=15, 50$ and $80$. And the rest three sub-figure are the Gaussian Retinex results with different $\sigma$.
• Figure 3: Architecture of SorroundNet. The cube presents the feature maps of the corresponding operation which is denoted by colorful rectangle, the detailed exhibition of ARBlocks and ECA can be found in Section. 3
• Figure 4: The visualization of building the ASF convolution kernel. In this toy example, $K$ is set to 5 and the normalization step is omitted.
• Figure 5: The architecture of ARBlock. The five images as shown below are the visualized results by averaging the channel of the feature maps.