arrow
返回

Scalable object instance recognition based on keygraph matching

delete2018-10-01
delete0
delete
OA
AI
E
Estephan Dazzi *
T
Teófilo de Campos
A
Adrian Hilton
R
Roberto M. César
DOI:10.1016/j.patrec.2017.10.038delete
delete原文链接
delete分享
delete收藏
查看原文
摘要

摘要

En 中文
We propose a generalisation of the local feature matching framework, where keypoints are replaced by k-keygraphs, i.e., isomorphic directed attributed graphs of cardinality k whose vertices are keypoints. Keygraphs have structural and topological properties which are discriminative and efficient to compute, based on graph edge length and orientation as well as vertex scale and orientation. Keypoint matching is performed based on descriptor similarity. Next, 2-keygraphs are calculated; as a result, the number of incorrect keypoint matches reduced in 75% (while the correct keypoint matches were preserved). Then, 3-keygraphs are calculated, followed by 4-keygraphs; this yielded a significant reduction of 99% in the number of remaining incorrect keypoint matches. The stage that finds 2-keygraphs has a computational cost equal to a small fraction of the cost of the keypoint matching stage, while the stages that find 3-keygraphs or 4-keygraphs have a negligible cost. In the final stage, RANSAC finds object poses represented as affine transformations mapping images. Our experiments concern large-scale object instance recognition subject to occlusion, background clutter and appearance changes. By using 4-keygraphs, RANSAC needed 1% of the iterations in comparison with 2-keygraphs or simple keypoints. As a result, using 4-keygraphs provided a better efficiency as well as allowed a larger number of initial keypoints matches to be established, which increased performance. (C) 2017 Elsevier B.V. All rights reserved.
Keyword:
Local invariant features
Semi-local graph matching
Graph topological properties
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

Pattern Recognition Letters 封面图
Pattern Recognition Letters
IF:
3.3
论文数:
8.0K
被引数:
2.0W

机构

U
University of Surrey
学者数:
662
论文数: 374
被引数: 9.5K
U
universidade de brasília
学者数:
289
论文数: 119
被引数: 2.7K
U
Universidade de São Paulo
学者数:
2.3K
论文数: 776
被引数: 3.0W
学者 查看更多机构
引用论文

引用论文

err分享
err收藏
A Comprehensive Performance Evaluation of 3D Local Feature Descriptors三维局部特征描述符的综合性能评价
err2015-04-16
err486
PREAI
errGuo, Yulan; Bennamoun, Mohammed; Sohel, Ferdous; Lu, Min; Wan, Jianwei; Kwok, Ngai Ming
err分享
err收藏
The Salem Smart Power Center: An Assessment of Battery Performance and Economic Potential
err
IF0
err2017-09-21
err0
errOAAI
errPatrick J. Balducci; M. J.E. Alam; Vilayanur Viswanathan; Di Wu; Aladsair J. Crawford; Kendall Mongird; Mark R. Weimar; Abhishek Somani; Kevin Whitener
err分享
err收藏
Saliency-based keypoint selection for fast object detection and matching
err2015-09-01
err21
PREAI
errBuoncompagni, Simone; Maio, Dario; Maltoni, Davide; Papi, Serena
err分享
err收藏
没有更多内容