# Real-World Single Image Super-Resolution Under Rainy Condition

**Authors:** Mohammad Shahab Uddin  
**arXiv:** [2206.08345](https://arxiv.org/abs/2206.08345)  
**Date:** 2022-06-16  
**Summary page:** https://www.summarizepaper.com/en/paper/2206.08345v1/

## Key points

- Image super-resolution is important in computer vision for various applications
- Real-world single image super-resolution is becoming popular due to its real-time application
- Challenging weather scenarios, such as rainy conditions, pose a challenge for image super-resolution
- Mohammad Shahab Uddin proposes a new algorithm using deep learning techniques to enhance the resolution of images under rainy conditions
- The proposed method outperforms existing state-of-the-art methods for real-world single image super-resolution under rainy conditions
- The approach involves training a neural network on pairs of low and high resolution images under different weather conditions, including rainy weather
- The proposed algorithm has potential applications in various fields such as surveillance and medical imaging.

## Summary

Image super-resolution is a crucial area of research in computer vision with numerous applications, including medical imaging and surveillance. Real-world single image super-resolution has become increasingly popular due to its real-time application. However, there are still many challenges to overcome, particularly during challenging weather scenarios such as rainy conditions. In this paper, Mohammad Shahab Uddin from Old Dominion University proposes a new algorithm for performing real-world single image super-resolution under rainy conditions. The proposed method mitigates the negative effects of rain on image super-resolution and improves the resolution of images from low to high resolution. The proposed algorithm uses deep learning techniques to enhance the resolution of images under rainy conditions. The approach involves training a neural network on pairs of low and high resolution images under different weather conditions, including rainy weather. The trained model can then be used to perform image super-resolution on new images captured in rainy conditions. To evaluate the effectiveness of the proposed algorithm, experiments were conducted using publicly available datasets containing images captured in rainy conditions. The results show that the proposed method outperforms existing state-of-the-art methods for real-world single image super-resolution under rainy conditions. In conclusion, this paper presents a novel approach for improving real-world single image super-resolution during challenging weather scenarios such as rain. The proposed algorithm demonstrates promising results and has potential applications in various fields such as surveillance and medical imaging. Future work could focus on extending this approach to other types of adverse weather conditions or exploring ways to improve the efficiency and speed of the algorithm for real time applications.