BasicSR
An open-source PyTorch toolkit for image and video restoration, with model training, super-resolution, denoising and deblurring support.
BasicSR is an image and video restoration toolkit for researchers and developers building software to recover detail or reduce visual defects. It provides restoration models and pipelines for training, testing and inference.
- Image restoration: Increase resolution, reduce noise, remove blur and clean up JPEG compression artifacts using models such as EDSR, RCAN, SRResNet, SRGAN, ESRGAN, SwinIR and ECBSR.
- Video super-resolution: Work with EDVR, BasicVSR and IconVSR, including bidirectional video super-resolution with BasicVSR and IconVSR.
- Model development: Train and test restoration models, run inference, and use Python package examples and templates as a starting point for your own projects.
- Quality evaluation: Assess image quality with NIQE and use plotting scripts to examine results.
BasicSR uses Python and PyTorch and is open source under the Apache 2.0 license. Its training and evaluation workflows suit people developing restoration models or incorporating them into software. It also supports StyleGAN2 and DFDNet. Real-ESRGAN uses BasicSR for general image restoration, while GFPGAN uses it for face restoration.
Logotyper
