Robust fitting: Assisted by semantic analysis and reinforcement learning

Author

Ha-Giang Truong, Edith Cowan UniversityFollow

Author Identifier

Giang Truong

https://orcid.org/0000-0002-7690-7739

Date of Award

2022

Document Type

Thesis

Publisher

Edith Cowan University

Degree Name

Doctor of Philosophy

School

School of Science

First Supervisor

David Suter

Second Supervisor

Syed Zulqarnain Gilani

Third Supervisor

Erchuan Zhang

Abstract

Many computer vision applications require robust model estimation from a set of observed data. However, these data usually contain outliers, due to imperfect data acquisition or pre-processing steps, which can reduce the performance of conventional model-fitting methods. Robust fitting is thus critical to make the model estimation robust against outliers and reach stable performance. All of the contributions made in this thesis are for maximum consensus. In robust model fitting, maximum consensus is one of the most popular criteria, which aims to estimate the model that is consistent to as many observations as possible, i.e. obtain the highest consensus. The thesis makes contributions in two aspects of maximum consensus, one is non-learning based approaches and the other is learning based approaches.

The first motivation for our work is the remarkable progress in semantic segmentation in recent years. Semantic segmentation is a useful process and is usually available for scene understanding, medical image analysis, and virtual reality. We propose novel methods, which make use of semantic segmentation, to improve the efficiency of two robust non-learning based algorithms.

Another motivation for our contributions is the advances in reinforcement learning. In the thesis, a novel unsupervised learning framework is proposed to learn (without labelled data) to solve robust estimation directly. In particular, we formulate robust fitting problem as a special case of goal-oriented learning, and adopt the Reinforcement Learning framework as the basis of our approach. Our approach is agnostic to the input features and can be generalized to various practical applications.

Related Publications

Truong, G., Gilani, S. Z., Islam, S. M. S., & Suter, D. (2019, December). Fast point cloud registration using semantic segmentation. In proceedings of the IEEE 2019 Digital Image Computing: Techniques and Applications (DICTA). IEEE. https://doi.org/10.1109/DICTA47822.2019.8945870

https://ro.ecu.edu.au/ecuworkspost2013/8146/

Giang Truong, Huu Le, Alvaro Parra, Syed Zulqarnain Gilani, Syed Islam, and David Suter (2020), Fast Semantic-Assisted Outlier Removal for Large-scale Point Cloud Registration. https://arxiv.org/abs/2202.10579

Truong, G., Le, H., Suter, D., Zhang, E., & Gilani, S. Z. (2021). Unsupervised learning for robust fitting: A reinforcement learning approach. In 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 10343-10352). IEEE. https://doi.org/10.1109/CVPR46437.2021.01021

https://ro.ecu.edu.au/ecuworkspost2013/11472/

G. Truong, H. Le, E. Zhang, D. Suter and S. Z. Gilani, (2023). Unsupervised learning for maximum consensus robust fitting: A reinforcement learning approach. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(3), 3890-3903. https://doi.org/10.1109/TPAMI.2022.3178442

https://ro.ecu.edu.au/ecuworks2022-2026/839/

Comments

Author also known as Giang Truong

Access Note

Access to this thesis is not available

Recommended Citation

Truong, H. (2022). Robust fitting: Assisted by semantic analysis and reinforcement learning. Edith Cowan University. Retrieved from https://ro.ecu.edu.au/theses/2567