jajatikr / video-inpainting Goto Github PK

Hi! I'm struggling a bit with running this as the input data is a bit vague. For now, I have created 50 frame .jpg images out of a video as an input in a "src-images" folder but the preprocessing files are causing issues (the "resize" and "video" files in "dataset" are creating errors). I believe this is caused by the input files being incorrect in some way.. would it be possible to post an example of a functioning input dataset? I am really interested in how this code functions and would love to see a functioning example.

Hi,
Thank for sharing your codes. I have a few question:
1- I am wondering if this code is an implementation of a published paper and if yes, could you please let me where can I read about the baseline?
2- Are you applying frame by frame image in-painting or you are considering the temporal aspect too?
3- Do you have any pre-trained model to test?

Thanks, Shek