Exploiting synthetic images for real-world image recognition

Bachelor Thesis (2018)
Author(s)

M.A.X. Maton (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Contributor(s)

Jan Van Gemert – Mentor

Miriam Huijser – Mentor

O.S. Kayhan – Mentor

Faculty
Electrical Engineering, Mathematics and Computer Science
Copyright
© 2018 Max Maton
More Info
expand_more
Publication Year
2018
Language
English
Copyright
© 2018 Max Maton
Graduation Date
02-07-2018
Awarding Institution
Delft University of Technology
Related content

Main research website

https://aiir.maxmaton.nl/

Main repository

https://github.com/thexa4/artificial-data-research
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

Creating big datasets is often difficult or expensive which causes people to augment their dataset with rendered images. This often fails to significantly improve accuracy due to a difference in distribution between real and rendered datasets. This paper shows that the gap between synthetic and real-world image distributions can be closed by using GANs to convert the synthetic data to a dataset which has the same distribution as the real data. Training this GAN requires only a fraction of the dataset traditionally required to get a high classification accuracy. This converted data can subsequently be used to train a classifier with a higher accuracy than a classifier trained only on the real dataset.

Files

Research.pdf
(pdf | 0.328 Mb)
License info not available