Classification Method of Image Feature Matching Using Naive Bayes Classifier

DOI

10.2991/asum.k.210827.058How to use a DOI?

Keywords

Feature Detection, Feature Matching, Heinly Dataset, Homography, Naive Bayes Classifier, ORB, PROSAC, VGG Dataset

Abstract

Since matching accuracy determines the performance of the overall algorithm, studies using image features require sophisticated classification technique for matching. However, there is a critical problem that the factors used to classify true or false matches are extremely limited. To solve this problem, we defined a new factor through geometric and statistical analysis of matched features. And then we performed the naive Bayes classifier with three factors to classify true or false matches. To verify the proposed method, we compared it with the traditional method using benchmark datasets (Heinly dataset, VGG dataset) where homography is provided as ground truth. As a result of the comparison experiments, the proposed method derived higher precision, recall, and F1 scores than the traditional method.

Copyright

© 2021, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Jin Yong Kim
AU  - Eun Kyeong Kim
AU  - Baekcheon Kim
AU  - Daekeon Ha
AU  - Sungshin Kim
PY  - 2021
DA  - 2021/08/30
TI  - Classification Method of Image Feature Matching Using Naive Bayes Classifier
BT  - Joint Proceedings of the 19th World Congress of the International Fuzzy Systems Association (IFSA), the 12th Conference of the European Society for Fuzzy Logic and Technology (EUSFLAT), and the 11th International Summer School on Aggregation Operators (AGOP)
PB  - Atlantis Press
SP  - 435
EP  - 442
SN  - 2589-6644
UR  - https://doi.org/10.2991/asum.k.210827.058
DO  - 10.2991/asum.k.210827.058
ID  - Kim2021
ER  -