Hierarchical progressive trust model for mismatch removal under both rigid and non-rigid transformations

Songlin Du, Takeshi Ikenaga

Research output: Contribution to journalArticle

Abstract

Accurate visual correspondence is the foundation of many computer vision based applications. Since existing image matching algorithms generate mismatches inevitably, a reliable mismatch-removal algorithm is highly desired to remove mismatches and preserve true matches. This paper proposes a hierarchical progressive trust (HPT) model to solve this problem. The HPT model first adopts a "trust the most trustworthy ones" strategy to select anchor inliers in its bottom layer, and then progressively propagates the trust from bottom layer to other layers in a bottomup way: 1) bottom layer verifies anchor inliers with the guidance of local features; 2) middle layers progressively estimate local transformations and perform local verifications; 3) top layer estimates a global transformation with an anchor-inliers-guided expectation maximization (EM) algorithm and performs global verifications. Experimental results show that the proposed HPT model achieves higher performance than state-of-the-art mismatch-removal methods under both rigid transformations and non-rigid deformations.

Original languageEnglish
Pages (from-to)1786-1794
Number of pages9
JournalIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
VolumeE101A
Issue number11
DOIs
Publication statusPublished - 2018 Nov 1

Fingerprint

Trust Model
Anchors
Image matching
Computer vision
Image Matching
Local Features
Expectation-maximization Algorithm
Matching Algorithm
Bottom-up
Estimate
Computer Vision
Guidance
Correspondence
High Performance
Verify
Experimental Results

Keywords

  • hierarchical progressive trust
  • image matching
  • mismatch removal
  • visual correspondence

ASJC Scopus subject areas

  • Signal Processing
  • Computer Graphics and Computer-Aided Design
  • Electrical and Electronic Engineering
  • Applied Mathematics

Cite this

@article{3bce0eac303342ada1163967f16c25e8,
title = "Hierarchical progressive trust model for mismatch removal under both rigid and non-rigid transformations",
abstract = "Accurate visual correspondence is the foundation of many computer vision based applications. Since existing image matching algorithms generate mismatches inevitably, a reliable mismatch-removal algorithm is highly desired to remove mismatches and preserve true matches. This paper proposes a hierarchical progressive trust (HPT) model to solve this problem. The HPT model first adopts a {"}trust the most trustworthy ones{"} strategy to select anchor inliers in its bottom layer, and then progressively propagates the trust from bottom layer to other layers in a bottomup way: 1) bottom layer verifies anchor inliers with the guidance of local features; 2) middle layers progressively estimate local transformations and perform local verifications; 3) top layer estimates a global transformation with an anchor-inliers-guided expectation maximization (EM) algorithm and performs global verifications. Experimental results show that the proposed HPT model achieves higher performance than state-of-the-art mismatch-removal methods under both rigid transformations and non-rigid deformations.",
keywords = "hierarchical progressive trust, image matching, mismatch removal, visual correspondence",
author = "Songlin Du and Takeshi Ikenaga",
year = "2018",
month = "11",
day = "1",
doi = "10.1587/transfun.E101.A.1786",
language = "English",
volume = "E101A",
pages = "1786--1794",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "11",

}

TY - JOUR

T1 - Hierarchical progressive trust model for mismatch removal under both rigid and non-rigid transformations

AU - Du, Songlin

AU - Ikenaga, Takeshi

PY - 2018/11/1

Y1 - 2018/11/1

N2 - Accurate visual correspondence is the foundation of many computer vision based applications. Since existing image matching algorithms generate mismatches inevitably, a reliable mismatch-removal algorithm is highly desired to remove mismatches and preserve true matches. This paper proposes a hierarchical progressive trust (HPT) model to solve this problem. The HPT model first adopts a "trust the most trustworthy ones" strategy to select anchor inliers in its bottom layer, and then progressively propagates the trust from bottom layer to other layers in a bottomup way: 1) bottom layer verifies anchor inliers with the guidance of local features; 2) middle layers progressively estimate local transformations and perform local verifications; 3) top layer estimates a global transformation with an anchor-inliers-guided expectation maximization (EM) algorithm and performs global verifications. Experimental results show that the proposed HPT model achieves higher performance than state-of-the-art mismatch-removal methods under both rigid transformations and non-rigid deformations.

AB - Accurate visual correspondence is the foundation of many computer vision based applications. Since existing image matching algorithms generate mismatches inevitably, a reliable mismatch-removal algorithm is highly desired to remove mismatches and preserve true matches. This paper proposes a hierarchical progressive trust (HPT) model to solve this problem. The HPT model first adopts a "trust the most trustworthy ones" strategy to select anchor inliers in its bottom layer, and then progressively propagates the trust from bottom layer to other layers in a bottomup way: 1) bottom layer verifies anchor inliers with the guidance of local features; 2) middle layers progressively estimate local transformations and perform local verifications; 3) top layer estimates a global transformation with an anchor-inliers-guided expectation maximization (EM) algorithm and performs global verifications. Experimental results show that the proposed HPT model achieves higher performance than state-of-the-art mismatch-removal methods under both rigid transformations and non-rigid deformations.

KW - hierarchical progressive trust

KW - image matching

KW - mismatch removal

KW - visual correspondence

UR - http://www.scopus.com/inward/record.url?scp=85056115085&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056115085&partnerID=8YFLogxK

U2 - 10.1587/transfun.E101.A.1786

DO - 10.1587/transfun.E101.A.1786

M3 - Article

AN - SCOPUS:85056115085

VL - E101A

SP - 1786

EP - 1794

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 11

ER -