Close menu


Sunderland Repository records the research produced by the University of Sunderland including practice-based research and theses.

Using outlier elimination to assess learning-based correspondence matching methods

Ding, Xintao, Luo, Yonglong, Jie, Biao, Li, Qingde and Cheng, Yongqiang (2024) Using outlier elimination to assess learning-based correspondence matching methods. Information Sciences, 659. ISSN 0020-0255

Item Type: Article


Recently, deep learning (DL) technology has been widely used in correspondence matching. The learning-based models are usually trained on benign image pairs with partial overlaps. Since DL model is usually data-dependent, non-overlapping images may be used as poison samples to fool the model and produce false registrations. In this study, we propose an outlier elimination based assessment method (OEAM) to assess the registrations of learning-based correspondence matching method on partially overlapping and non-overlapping images. OEAM first eliminates outliers based on spatial paradox. Then OEAM implements registration assessment in two streams using the obtained core correspondence set. If the cardinality of the core set is sufficiently small, the input registration is assessed as a low-quality registration. Otherwise, it is assessed to be of high quality, and OEAM improves its registration performance using the core set. OEAM is a post-processing technique imposed on learning-based method. The comparison experiments are implemented on outdoor (YFCC100M) and indoor (SUN3D) datasets using four deep learning-based methods. The experimental results on registrations of partially overlapping images show that OEAM can reliably infer low-quality registrations and improve performance on high-quality registrations. The experiments on registrations of non-overlapping images demonstrate that learning-based methods are vulnerable to poisoning attacks launched by non overlapping images, and OEAM is robust against poisoning attacks crafted by non-overlapping images.

INS-D-23-5555-R2.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (26MB) | Preview

More Information

Depositing User: Yongqiang Cheng


Item ID: 17204
Identification Number:
ISSN: 0020-0255
Official URL:

Users with ORCIDS

ORCID for Xintao Ding: ORCID iD
ORCID for Yonglong Luo: ORCID iD
ORCID for Qingde Li: ORCID iD
ORCID for Yongqiang Cheng: ORCID iD

Catalogue record

Date Deposited: 11 Jan 2024 12:22
Last Modified: 11 Jan 2024 12:30


Author: Xintao Ding ORCID iD
Author: Yonglong Luo ORCID iD
Author: Qingde Li ORCID iD
Author: Yongqiang Cheng ORCID iD
Author: Biao Jie

University Divisions

Faculty of Technology > School of Computer Science


Computing > Data Science
Computing > Artificial Intelligence

Actions (login required)

View Item (Repository Staff Only) View Item (Repository Staff Only)