Automatic keypoint detecting of wireframe gates

Master Thesis (2019)
Author(s)

Wei Sun (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Contributor(s)

J.C. Gemert – Mentor (TU Delft - Pattern Recognition and Bioinformatics)

Faculty
Electrical Engineering, Mathematics and Computer Science
Copyright
© 2019 Wei Sun
More Info
expand_more
Publication Year
2019
Language
English
Copyright
© 2019 Wei Sun
Graduation Date
11-07-2019
Awarding Institution
Delft University of Technology
Programme
['Electrical Engineering | Embedded Systems']
Faculty
Electrical Engineering, Mathematics and Computer Science
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

This work applies keypoint detection method to solve gate recognition problem. Unlike regular object detection task, gate recognition problem is made difficult by the fact that gate is empty wireframe which means that the object surrounded by gate-edge is not relevant and should not be taken into consideration when detecting. However, regular object detection algorithms will process on whole pixels of specific region and give the results as bounding box with object class. The architecture used in this project consists of two branches which are corner detector and edge detector respectively. Detected corners and edges are highlighted in heatmaps. We first verify the correctness of our model in the toy dataset and upgrade the model to work on more complex dataset. The experimental evidence shows the performance and functionality of our network intuitively.

Files

Thesis_final.pdf
(pdf | 3.02 Mb)
License info not available