Awesome

NEST: Neural Event Stack for Event-based Image Enhancement, ECCV 2022

Minggui Teng, Chu Zhou, Hanyue Lou, and Boxin Shi.

network

Dependency

pip install -r requirements.txt

Train

Deblur

python train.py --mode deblur --TrainImgPath `Path of Train Image` --TrainEvePath `Path of Events` --TrainGTPath `Path of Ground Truth Image`

SR

python train.py --mode sr --upsample_scale 4 --TrainImgPath `Path of Train Image` --TrainEvePath `Path of Events` --TrainGTPath `Path of Ground Truth Image`

Test

Pretrained model can be download in PKU Disk or Google Drive, and copy to pretrained file folder.

Deblur

python test.py --mode deblur --ckp pretrained/model_deblur_best.pth --TestImgPath demo_input/blur --TestEvePath demo_input/events

SR

python test.py --mode sr --upsample_scale 4 --ckp pretrained/model_sr_best.pth --TestImgPath demo_input/lr --TestEvePath demo_input/events

configurations can be changed in utils/options.py

Examples of HFR video generation application

examples1

More visual results are in the subfolder demo.

Contact

If you have any questions, please send an email to minggui_teng@pku.edu.cn