Learnable Graph Matching: A Practical Paradigm for Data Association

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: He, Jiawei, Huang, Zehao, Wang, Naiyan, Zhang, Zhaoxiang
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913224002109440
author He, Jiawei
Huang, Zehao
Wang, Naiyan
Zhang, Zhaoxiang
author_facet He, Jiawei
Huang, Zehao
Wang, Naiyan
Zhang, Zhaoxiang
contents Data association is at the core of many computer vision tasks, e.g., multiple object tracking, image matching, and point cloud registration. however, current data association solutions have some defects: they mostly ignore the intra-view context information; besides, they either train deep association models in an end-to-end way and hardly utilize the advantage of optimization-based assignment methods, or only use an off-the-shelf neural network to extract features. In this paper, we propose a general learnable graph matching method to address these issues. Especially, we model the intra-view relationships as an undirected graph. Then data association turns into a general graph matching problem between graphs. Furthermore, to make optimization end-to-end differentiable, we relax the original graph matching problem into continuous quadratic programming and then incorporate training into a deep graph neural network with KKT conditions and implicit function theorem. In MOT task, our method achieves state-of-the-art performance on several MOT datasets. For image matching, our method outperforms state-of-the-art methods on a popular indoor dataset, ScanNet. For point cloud registration, we also achieve competitive results. Code will be available at https://github.com/jiaweihe1996/GMTracker.
format Preprint
id arxiv_https___arxiv_org_abs_2303_15414
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Learnable Graph Matching: A Practical Paradigm for Data Association
He, Jiawei
Huang, Zehao
Wang, Naiyan
Zhang, Zhaoxiang
Computer Vision and Pattern Recognition
Data association is at the core of many computer vision tasks, e.g., multiple object tracking, image matching, and point cloud registration. however, current data association solutions have some defects: they mostly ignore the intra-view context information; besides, they either train deep association models in an end-to-end way and hardly utilize the advantage of optimization-based assignment methods, or only use an off-the-shelf neural network to extract features. In this paper, we propose a general learnable graph matching method to address these issues. Especially, we model the intra-view relationships as an undirected graph. Then data association turns into a general graph matching problem between graphs. Furthermore, to make optimization end-to-end differentiable, we relax the original graph matching problem into continuous quadratic programming and then incorporate training into a deep graph neural network with KKT conditions and implicit function theorem. In MOT task, our method achieves state-of-the-art performance on several MOT datasets. For image matching, our method outperforms state-of-the-art methods on a popular indoor dataset, ScanNet. For point cloud registration, we also achieve competitive results. Code will be available at https://github.com/jiaweihe1996/GMTracker.
title Learnable Graph Matching: A Practical Paradigm for Data Association
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2303.15414