Working hard to know your neighbor's margins: Local descriptor learning loss

Anastasiya Mishchuk, Dmytro Mishkin, Filip Radenovic, Jiri Matas
5/30/2017
cs.CV

Official code: DagnyT/hardnet (534 stars, MIT, PyTorch, last commit 2025-04-14).

Abstract

We introduce a novel loss for learning local feature descriptors which is inspired by the Lowe's matching criterion for SIFT. We show that the proposed loss that maximizes the distance between the closest positive and closest negative patch in the batch is better than complex regularization methods; it works well for both shallow and deep convolution network architectures. Applying the novel loss to the L2Net CNN architecture results in a compact descriptor -- it has the same dimensionality as SIFT (128) that shows state-of-art performance in wide baseline stereo, patch verification and instance retrieval benchmarks. It is fast, computing a descriptor takes about 1 millisecond on a low-end GPU.

View on arXivView PDF

Code Implementations(1)

DagnyT/hardnetOfficial100%
PyTorchMIT

Hardnet descriptor model - "Working hard to know your neighbor's margins: Local descriptor learning loss"

534101Shell, PythonMay 29, 20171 years agoMIT
computer-visionconvolutional-neural-networksdeep-learningimage-matchingimage-retrieval+3 more

Similar Papers

Learning Spread-out Local Feature Descriptors

86% match1 repo412017

Improving the HardNet Descriptor

84% match1 repo262020

Residual Learning for Image Point Descriptors

82% match2023

A Sliced Wasserstein Loss for Neural Texture Synthesis

80% match1 repo612020

Support Vector Guided Softmax Loss for Face Recognition

80% match1 repo1862018

S-TREK: Sequential Translation and Rotation Equivariant Keypoints for local feature extraction

80% match2023

Cite this paper

@article{mishchuk2017working,
  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},
  year   = {2017},
  eprint = {1705.10872},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1705.10872}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1705.10872.svg)](https://codeofpaper.com/paper/1705.10872)

Discussion