Summary of Rsend: Retinex-based Squeeze and Excitation Network with Dark Region Detection For Efficient Low Light Image Enhancement, by Jingcheng Li et al.
RSEND: Retinex-based Squeeze and Excitation Network with Dark Region Detection for Efficient Low Light Image Enhancement
by Jingcheng Li, Ye Qiao, Haocheng Xu, Sitao Huang
First submitted to arxiv on: 14 Jun 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents RSEND, a novel framework for enhancing low-light images using Retinex theory. Previous CNN-based methods struggled with complex datasets like LOL-v2, consuming excessive computational resources and requiring multi-stage training. In contrast, RSEND is a one-stage, concise approach that divides the image into illumination and reflectance maps, captures details in the illumination map, enhances gray-scale images, and refines the output using element-wise matrix multiplication and denoising. The Squeeze and Excitation network is employed throughout to better capture details. Experimental results demonstrate RSEND’s superiority over CNN-based models (0.44 dB to 4.2 dB PSNR improvement) and transformer-based models in certain datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper solves a big problem with low-light images: they’re often blurry and hard to use. The authors created a new way to make these images clearer using an old idea called Retinex theory. Their method, RSEND, is fast and efficient, unlike other methods that take a long time to train. It works by separating the image into two parts (the light and what’s being lit), making the light part brighter, and then combining the two parts again. This makes the final image much clearer than before. The authors tested their method and found it was better than others at improving low-light images. |
Keywords
* Artificial intelligence * Cnn * Transformer
