-
Create a local copy of repository using the following commands
foor@bar:~$ git clone https://github.com/sanghviyashiitb/structured-kernel-cvpr23.git foor@bar:~$ cd structured-kernel-cvpr23 foor@bar:~/structured-kernel-cvpr23$
-
Download the pretrained models, i.e. denoiser, p4ip, and ktn into
model_zoo
from the link here -
To test the network on levin-data, run the file
foor@bar:~/structured-kernel-cvpr23$ python3 demo_grayscale.py
-
To test the network on real-sensor data, run the file
foor@bar:~/structured-kernel-cvpr23$ python3 demo_real.py
For further details on this dataset containing real-sensor noise + motion blur along with ground-truth kernels i.e., Photon Limited Deblurring Dataset (PLDD) refer to this link
@InProceedings{Sanghvi_2023_CVPR,
author = {Sanghvi, Yash and Mao, Zhiyuan and Chan, Stanley H.},
title = {Structured Kernel Estimation for Photon-Limited Deconvolution},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2023},
pages = {9863-9872}
}
Feel free to ask your questions/share your feedback using the Issues feature at the the top of this page.