BSRGAN
Practical image super-resolution model trained on diverse degradations.
About
BSRGAN from the Computer Vision Lab at ETH Zurich is a practical image super-resolution model trained with a degradation pipeline that randomly combines blur, noise, JPEG compression, and resizing, so it generalizes to real-world low-quality images better than models trained on simple synthetic downsampling. It includes training and test code and pretrained weights. Released under the Apache 2.0 license.
Reviews (0)
Leave a Review
No reviews yet. Be the first to review!
Details
- Category
- AI Image/Video Editing
- Price
- Free
- Platform
- Local/Desktop
- Difficulty
- Easy (2/5)
- License
- Apache-2.0
- Minimum VRAM
- 4 GB
- Added
- Apr 3, 2026
Related Tools
Image restoration using Swin Transformer for super-resolution and denoising.
AI-powered image inpainting tool for removing objects and defects from photos.
Modular Stable Diffusion web UI focused on power users and server deployment.
Face restoration model by Tencent ARC for enhancing old or low-quality face photos.
Optimized fork of AUTOMATIC1111 with better memory management and speed.
Robust face restoration using codebook lookup transformer.