Semantic matching by weakly supervised 2D point set registration

Zakaria Laskar, Hamed R. Tavakoli, Juho Kannala

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

5 Citations (Scopus)
245 Downloads (Pure)

Abstract

In this paper we address the problem of establishing correspondences between different instances of the same object. The problem is posed as finding the geometric transformation that aligns a given image pair. We use a convolutional neural network (CNN) to directly regress the parameters of the transformation model. The alignment problem is defined in the setting where an unordered set of semantic key-points per image are available, but, without the correspondence information. To this end we propose a novel loss function based on cyclic consistency that solves this 2D point set registration problem by inferring the optimal geometric transformation model parameters. We train and test our approach on a standard benchmark dataset Proposal-Flow (PF-PASCAL)[8]. The proposed approach achieves state-of-the-art results demonstrating the effectiveness of the method. In addition, we show our approach further benefits from additional training samples in PF-PASCAL generated by using category level information.

Original languageEnglish
Title of host publication2019 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV)
PublisherIEEE
Pages1061-1069
Number of pages9
ISBN (Electronic)9781728119755
DOIs
Publication statusPublished - 4 Mar 2019
MoE publication typeA4 Conference publication
EventIEEE Winter Conference on Applications of Computer Vision - Waikoloa Village, United States
Duration: 7 Jan 201911 Jan 2019
Conference number: 19

Conference

ConferenceIEEE Winter Conference on Applications of Computer Vision
Abbreviated titleWACV
Country/TerritoryUnited States
CityWaikoloa Village
Period07/01/201911/01/2019

Fingerprint

Dive into the research topics of 'Semantic matching by weakly supervised 2D point set registration'. Together they form a unique fingerprint.

Cite this