Annotation for the frame field

[kriti115]

I have the angles for each pixel as a .npy file to be used as groundtruth as mentioned in the paper.

My question is: is this directly used in the network or is there a need to calculate frame field to be fed into the network? If so what should the format be?

I would appreciate any kind of help. Thank you.

[patriksabol]

The angles to be fed into the network are generated during training data generation process implemented as data_transforms (f.e. https://github.com/Lydorn/Polygonization-by-Frame-Field-Learning/blob/master/dataset_folds.py). Your data should be located in "raw" folder, after that, training data are located in "processed" folder. You just need to adjust transformation to your own data.

[kriti115]

@patriksabol Thank you for you input. I was able to understand that we only need the images and binary masks as input and not the angles as they will be calculated while running the network.

I was able to run the network and got the training data as '.pt' file in processed folder. Although the training stopped with error half way I was still able to get half of the data in processed folder.

I will close this issue now.