use pretrain model to deblur my images

[SYuan-X]

Hi, author. Your work is really excellent, and I am particularly interested in applying the code provided in the paper to my own set of images by using your pretrained model. Could you please provide some guidance about how to convert my own data into the format of input data and use the pretrained model to deblur?

Thank you very much for your time and consideration.

[XiangZ-0]

Hi, thanks for your interest in our work. The training and testing data formats are the same in our works. I recommend converting your data to the format like our MS-RBD dataset (you can also check the npz files in our released datasets for details), and then modify the configuration files here, e.g., msrbd_test.yaml. For testing, you only need to modify the following parameters according to your settings:

load_dir: checkpoint/MS-RBD # change it to your path to load checkpoints
root_path: datasets/MS-RBD # change it to your dataset directory
save_path: results/MS-RBD # change it to your result directory
scale_factor: 4 # change it according to the spatial resolution ratio of images over events in your dataset

Then it should work. Hope this helps :)