Restormer-Plus for Real World Image Deraining: One State-of-the-Art Solution to the GT-RAIN Challenge (CVPR 2023 UG2+ Track 3)

Chaochao Zheng,Luping Wang,Bin Liu

2024-10-11

Abstract:This technical report presents our Restormer-Plus approach, which was submitted to the GT-RAIN Challenge (CVPR 2023 UG$^2$+ Track 3). Details regarding the challenge are available at <a class="link-external link-http" href="http://cvpr2023.ug2challenge.org/track3.html" rel="external noopener nofollow">this http URL</a>. Restormer-Plus outperformed all other submitted solutions in terms of peak signal-to-noise ratio (PSNR), and ranked 4th in terms of structural similarity (SSIM). It was officially evaluated by the competition organizers as a runner-up solution. It consists of four main modules: the single-image de-raining module (Restormer-X), the median filtering module, the weighted averaging module, and the post-processing module. Restormer-X is applied to each rainy image and built on top of Restormer. The median filtering module is used as a median operator for rainy images associated with each scene. The weighted averaging module combines the median filtering results with those of Restormer-X to alleviate overfitting caused by using only Restormer-X. Finally, the post-processing module is utilized to improve the brightness restoration. These modules make Restormer-Plus one of the state-of-the-art solutions for the GT-RAIN Challenge. Our code can be found at <a class="link-external link-https" href="https://github.com/ZJLAB-AMMI/Restormer-Plus" rel="external noopener nofollow">this https URL</a>.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper attempts to address the problem of improving the quality of image restoration in real-world image deraining tasks. Specifically, the paper focuses on the following issues:

1. **Insufficient detail and texture recovery capability**: Existing deraining methods often struggle to restore clear details and textures when processing rainy images.
2. **Overfitting during the training process**: During training, models tend to overfit the training data, leading to poor performance on the test set.
3. **Suboptimal brightness recovery**: Existing methods have shortcomings in restoring image brightness, resulting in a significant difference between the restored image brightness and the actual non-rainy image.

To address these issues, the authors propose a method called Restormer-Plus, which includes improvements in several modules, such as the single image deraining module, median filtering module, weighted averaging module, and post-processing module. With these improvements, Restormer-Plus achieves significant enhancements in metrics such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM), and has achieved excellent results in the GT-RAIN challenge.