Two-Pass K Nearest Neighbor Search for Feature Tracking

In recent years, feature tracking has become one of the most important research topics in computer vision. Many efforts have been made to design excellent feature matching methods. For large-scale structure from motion, however, existing feature tracking methods still need to improve in aspects of s...

Full description

Saved in:
Bibliographic Details
Main Authors: Mingwei Cao, Wei Jia, Zhihan Lv, Wenjun Xie, Liping Zheng, Xiaoping Liu
Format: Article
Language:English
Published: IEEE 2018-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/8528423/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849731610627276800
author Mingwei Cao
Wei Jia
Zhihan Lv
Wenjun Xie
Liping Zheng
Xiaoping Liu
author_facet Mingwei Cao
Wei Jia
Zhihan Lv
Wenjun Xie
Liping Zheng
Xiaoping Liu
author_sort Mingwei Cao
collection DOAJ
description In recent years, feature tracking has become one of the most important research topics in computer vision. Many efforts have been made to design excellent feature matching methods. For large-scale structure from motion, however, existing feature tracking methods still need to improve in aspects of speed and matching confidence. To defense the drawbacks, in this paper, we design a simple and efficient feature tracking method based on the standard <inline-formula> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> nearest neighbor search. First, the parallel scale-invariant feature transform (SIFT) is selected as the feature detector to locate keypoints. Second, the principal component analysis-based SIFT-descriptor extractor is used to compute robust descriptions for the selected keypoints, in which normalized operation is used for boosting the matching score. Third, the two-pass <inline-formula> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> nearest neighbor search (TP-KNN) is proposed to produce correspondences for image pairs, then leading a significant improvement in the number of matches. Moreover, a geometry-constraint approach is proposed to remove outliers from the initial matches for boosting the matching precision. Finally, we conduct experiment on several challenging benchmark datasets to assess the TP-KNN method against the state-of-the-art methods. Experimental results indicate that the TP-KNN has the best performance in both speed and accuracy.
format Article
id doaj-art-aef75f2976c0423c8375d8fcc48b53ae
institution DOAJ
issn 2169-3536
language English
publishDate 2018-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-aef75f2976c0423c8375d8fcc48b53ae2025-08-20T03:08:31ZengIEEEIEEE Access2169-35362018-01-016729397295110.1109/ACCESS.2018.28793378528423Two-Pass K Nearest Neighbor Search for Feature TrackingMingwei Cao0Wei Jia1Zhihan Lv2Wenjun Xie3https://orcid.org/0000-0002-4032-8049Liping Zheng4Xiaoping Liu5School of Computer Science and Information Engineering, Hefei University of Technology, Hefei, ChinaSchool of Computer Science and Information Engineering, Hefei University of Technology, Hefei, ChinaSchool of Data Science and Software Engineering, Qingdao University, Qingdao, ChinaSchool of Computer Science and Information Engineering, Hefei University of Technology, Hefei, ChinaSchool of Computer Science and Information Engineering, Hefei University of Technology, Hefei, ChinaSchool of Computer Science and Information Engineering, Hefei University of Technology, Hefei, ChinaIn recent years, feature tracking has become one of the most important research topics in computer vision. Many efforts have been made to design excellent feature matching methods. For large-scale structure from motion, however, existing feature tracking methods still need to improve in aspects of speed and matching confidence. To defense the drawbacks, in this paper, we design a simple and efficient feature tracking method based on the standard <inline-formula> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> nearest neighbor search. First, the parallel scale-invariant feature transform (SIFT) is selected as the feature detector to locate keypoints. Second, the principal component analysis-based SIFT-descriptor extractor is used to compute robust descriptions for the selected keypoints, in which normalized operation is used for boosting the matching score. Third, the two-pass <inline-formula> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> nearest neighbor search (TP-KNN) is proposed to produce correspondences for image pairs, then leading a significant improvement in the number of matches. Moreover, a geometry-constraint approach is proposed to remove outliers from the initial matches for boosting the matching precision. Finally, we conduct experiment on several challenging benchmark datasets to assess the TP-KNN method against the state-of-the-art methods. Experimental results indicate that the TP-KNN has the best performance in both speed and accuracy.https://ieeexplore.ieee.org/document/8528423/Feature trackingK nearest neighbor3D reconstructionstructure from motionfeature matching
spellingShingle Mingwei Cao
Wei Jia
Zhihan Lv
Wenjun Xie
Liping Zheng
Xiaoping Liu
Two-Pass K Nearest Neighbor Search for Feature Tracking
IEEE Access
Feature tracking
K nearest neighbor
3D reconstruction
structure from motion
feature matching
title Two-Pass K Nearest Neighbor Search for Feature Tracking
title_full Two-Pass K Nearest Neighbor Search for Feature Tracking
title_fullStr Two-Pass K Nearest Neighbor Search for Feature Tracking
title_full_unstemmed Two-Pass K Nearest Neighbor Search for Feature Tracking
title_short Two-Pass K Nearest Neighbor Search for Feature Tracking
title_sort two pass k nearest neighbor search for feature tracking
topic Feature tracking
K nearest neighbor
3D reconstruction
structure from motion
feature matching
url https://ieeexplore.ieee.org/document/8528423/
work_keys_str_mv AT mingweicao twopassknearestneighborsearchforfeaturetracking
AT weijia twopassknearestneighborsearchforfeaturetracking
AT zhihanlv twopassknearestneighborsearchforfeaturetracking
AT wenjunxie twopassknearestneighborsearchforfeaturetracking
AT lipingzheng twopassknearestneighborsearchforfeaturetracking
AT xiaopingliu twopassknearestneighborsearchforfeaturetracking