https://github.com/ducha-aiki/hesaff-pytorch
PyTorch implementation of Hessian-Affine local feature detector
https://github.com/ducha-aiki/hesaff-pytorch
cnn convolutional-neural-networks deep-learning detector hessian hessian-affine local-features pytorch
Last synced: 3 months ago
JSON representation
PyTorch implementation of Hessian-Affine local feature detector
- Host: GitHub
- URL: https://github.com/ducha-aiki/hesaff-pytorch
- Owner: ducha-aiki
- Created: 2017-10-12T15:06:22.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2017-11-22T10:45:36.000Z (over 8 years ago)
- Last Synced: 2025-03-30T22:11:15.354Z (about 1 year ago)
- Topics: cnn, convolutional-neural-networks, deep-learning, detector, hessian, hessian-affine, local-features, pytorch
- Language: Python
- Size: 58.6 KB
- Stars: 26
- Watchers: 5
- Forks: 4
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
This is implementation of Hessian-Affine local feature detector.
It is heavily based on Michal Perdoch C++ implementation https://github.com/perdoch/hesaff
pytaff - current implementation
hesamp - Michal Perdoch C++ one.
There are several differences:
1) No SIFT description, the output is image patches. If one needs to, patches could be feed into [PyTorchSIFT](https://github.com/ducha-aiki/pytorch-sift)
2) Subpixel precision is done via "center-of-responce-mass" inspired by [LIFT](https://arxiv.org/abs/1603.09114) paper, instead of original iterative quadratic fitting
3) Instead of setting threshold to control number of detection, this implementation simply outputs top-K local extreme points.
You also might be interested in [HesAffNet](https://github.com/ducha-aiki/affnet), which gives significantly better results because of learned affine shape estimation procedure.
If you use this code for academic purposes, please cite the following paper:
```
@article{AffNet2017,
author = {Dmytro Mishkin, Filip Radenovic, Jiri Matas},
title = "{Learning Discriminative Affine Regions via Discriminability}",
year = 2017,
month = nov}
```