Robust point matching via spatial clustering and Gaussian mixture models

Xin Hu; Diyuan Xiang; Lijun Ma; Hao Qin

doi:10.1117/12.2658890

23 November 2022 Robust point matching via spatial clustering and Gaussian mixture models

Xin Hu, Diyuan Xiang, Lijun Ma, Hao Qin

Proceedings Volume 12454, International Symposium on Robotics, Artificial Intelligence, and Information Engineering (RAIIE 2022); 124540B (2022) https://doi.org/10.1117/12.2658890
Event: International Symposium on Robotics, Artificial Intelligence, and Information Engineering (RAIIE 2022), 2022, Hohhot, China

Abstract

In this paper, combine spatial clustering and Gaussian mixture models to propose an efficient and robust point set alignment scheme. This work aims to balance recall, accuracy, robustness and operational efficiency. First, the point set alignment problem is transformed into a spatial clustering problem with outliers. The main idea is to adaptively cluster the assumed matches into multiple motion-consistent clusters and one outlier cluster. The Gaussian mixture model is then fitted to each of the multiple kinematic consistency clusters. In order to pursue a higher alignment, the mixing coefficients of the Gaussian mixture model are assigned values using descriptors. This scheme is suitable for alignment problems with complex point sets. In addition, it offers better robustness and efficiency compared to existing related schemes. Analysis of the combined performance and simulation results demonstrates the effectiveness and robustness of the scheme.

Citation Download Citation

Xin Hu, Diyuan Xiang, Lijun Ma, and Hao Qin "Robust point matching via spatial clustering and Gaussian mixture models", Proc. SPIE 12454, International Symposium on Robotics, Artificial Intelligence, and Information Engineering (RAIIE 2022), 124540B (23 November 2022); https://doi.org/10.1117/12.2658890

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available