ku-cvlab/nemf

Official code implementation of NeMF (NeurIPS'22)

nemf's Issues

Visual Correspondence Matching

Hi,

Enjoyed looking over the paper. I was wondering if you could comment at all on the suitability of this work for visual (rather than semantic) correspondence matching? I.e. For use in an image stitching, or structure-from-motion context.

Link for the pre-trained weights is broken

The link to the google drive shows Error 404.
Could you fix it so that I can run your model with pretrained weights?

Possible bug in utils_training/evaluation.py

I implemented the training code myself, but my trained model yielded PCK less than 10 at the threshold 0.1.
Then, I found that the function, eval_kps_transfer_test() is using source keypoints as ground truth instead of target keypoints.
After the fix, I get a PCK of about 42 at the threshold 0.1

Inside the function eval_kps_transfer_test(cls, prd_kps, batch, alpha), batch['src_kps'] should be changed to batch['trg_kps']

Could you verify whether this is right?

Thank you.

When to open source training code

Hello, your work is very good.
When can the training code be open source?