.png){kind=icon}
返回
Exploiting multi-layer graph factorization for multi-attributed graph matching
.png){kind=icon}
OA
AI
DOI：10.1016/j.patrec.2018.09.024.png)
.png){kind=icon}
.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Multi-layer graph matching methods effectively solve multi-attributed graph matching problems based on the multi-layer structure adopted to address the ambiguity and uncertainty arisen from the attribute integration. However, despite of its effectiveness for matching multi-attributed graphs, the approach has a long way to apply in the practical environment due to the scalability problem caused by a huge matrix for describing the multi-layer structure. In this paper, we propose a novel multi-layer graph matching algorithm based on the multi-layer graph factorization to address the issue. The main contribution of this research is three-fold. First, we propose a factorization method that decomposes the huge multi-layer matrix into several small matrices for efficiency. Second, we reformulate the original multi-layer matching problem into two relaxed problems by using the factorized matrices. Third, based on the relaxed problems, we propose a multi-layer graph matching algorithm inspired from the convex-concave relaxation procedure. In our extensive experiments on the synthetic and real-image datasets, the proposed method exhibits better performance than state-of-the-art algorithms based on the single-layer structure. (C) 2018 Elsevier B.V. All rights reserved.
Keyword:
Multi-attributed graph matching
Multi-layer structure
Matrix factorization
Path following
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
IF:
3.3
论文数:
7.9K
被引数:
1.6W
机构
暂无机构信息
引用论文
Managing Service Productivity In Sub-saharan African Banks: The Case of Nigerian Knowledge Workers
PREAI
分享收藏Approximate graph edit distance computation by means of bipartite graph matching通过二分图匹配计算近似图编辑距离
PREAI
分享收藏Modular DNA-origami-based nanoarrays enhanced cell binding-affinity through “lock-key” interaction.png)
OAAI
IF0
OAAI分享收藏
查看更多内容