search

eric-yyjau / pytorch-superpoint

Superpoint Implemented in PyTorch: https://arxiv.org/abs/1712.07629
MIT License
780 stars 164 forks source link

Rotation #80

Open adamas-v opened 2 years ago

adamas-v commented 2 years ago

Nice work, but I visualized the matching of Hpatches and superpoint failed in all rotation case.maybe rotation not include in training augmentation?

ghost commented 1 year ago

the matching of Hpatches and SuperPoint failed in all rotation case Me too. My bad case of the result was only two lines in point matching of Hpatches. I don't know why. Probably, Some of the steps may be lacking in these codes.

adamas-v commented 1 year ago

the matching of Hpatches and SuperPoint failed in all rotation case Me too. My bad case of the result was only two lines in point matching of Hpatches. I don't know why. Probably, Some of the steps may be lacking in these codes.

that's beacuse the pre-train model was trained with little sample of rotation, so it is leak of ratotion attack .

ghost commented 1 year ago

that's beacuse the pre-train model was trained with little sample of rotation, so it is leak of ratotion attack .

Oh, I see. To deal with problem, it will be better to increase the rotation samples to train the pre-trained model. These codes didn't mention the step which is training MagicPoint with MS-COCO after Homographic Adaptation. These things may connect to this problem.