SDI-Net: Toward Sufficient Dual-View Interaction for Low-light Stereo Image Enhancement

Linlin Hu,Ao Sun,Shijie Hao,Richang Hong,Meng Wang

2024-08-20

Abstract:Currently, most low-light image enhancement methods only consider information from a single view, neglecting the correlation between cross-view information. Therefore, the enhancement results produced by these methods are often unsatisfactory. In this context, there have been efforts to develop methods specifically for low-light stereo image enhancement. These methods take into account the cross-view disparities and enable interaction between the left and right views, leading to improved performance. However, these methods still do not fully exploit the interaction between left and right view information. To address this issue, we propose a model called Toward Sufficient Dual-View Interaction for Low-light Stereo Image Enhancement (SDI-Net). The backbone structure of SDI-Net is two encoder-decoder pairs, which are used to learn the mapping function from low-light images to normal-light images. Among the encoders and the decoders, we design a module named Cross-View Sufficient Interaction Module (CSIM), aiming to fully exploit the correlations between the binocular views via the attention mechanism. The quantitative and visual results on public datasets validate the superiority of our method over other related methods. Ablation studies also demonstrate the effectiveness of the key elements in our model.

Computer Vision and Pattern Recognition,Artificial Intelligence,Image and Video Processing

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve

This paper aims to address the issue of insufficient dual-view information interaction in Low-Light Stereo Image Enhancement (LLSIE). Specifically:

1. **Limitations of Existing Methods**:
- Most current low-light image enhancement methods only consider information from a single view, neglecting the correlation between cross-view information, which often leads to unsatisfactory enhancement results.
- Although there are existing methods for low-light stereo image enhancement, these methods still lack sufficient interaction between the left and right views.

2. **Proposed New Method**:
- A model named SDI-Net (Sufficient Dual-View Interaction for Low-light Stereo Image Enhancement) is proposed to improve the enhancement effect of low-light stereo images by establishing sufficient interaction between the left and right views.
- SDI-Net employs two identical encoder-decoder structures and designs a module named Cross-View Sufficient Interaction Module (CSIM) to enhance the interaction between the features of the left and right views.
- The CSIM module includes the Cross-View Attention Interaction Module (CAIM) and the Pixel and Channel Attention Block (PCAB), which model attention mechanisms from the view level, channel level, and pixel level, respectively.

With these improvements, SDI-Net can better restore image details and textures under low-light conditions, and experiments on public datasets have validated its superiority and effectiveness.