kornia/hardnet

Pretrained weights for HardNet and HardNet++, used by kornia.feature.HardNet.

HardNet is a 128-dimensional descriptor for 32×32 grayscale patches, trained with a hard-negative mining triplet loss. NeurIPS 2017.

Original repo: DagnyT/hardnet

Weights

File	Training data
HardNetPP.pth	All Brown datasets (HardNet++)
checkpoint_liberty_with_aug.pth	Liberty with photometric augmentation
hardnet8v2.pt	HardNet8 v2, trained on HPatches

Citation

@inproceedings{HardNet2017,
    title     = {Working hard to know your neighbor's margins:
                 Local descriptor learning loss},
    author    = {Anastasiya Mishchuk and Dmytro Mishkin
                 and Filip Radenovic and Jiri Matas},
    booktitle = {NeurIPS},
    year      = {2017}
}

@article{HardNet2020,
    title   = {Improving the HardNet Descriptor},
    author  = {Milan Pultar},
    journal = {arXiv ePrint 2007.09699},
    year    = {2020}
}