To read the full version of this content please select one of the options below:

Image editing-based data augmentation for illumination-insensitive background subtraction

Dimitrios Sakkos (Department of Computer and Information Sciences, Northumbria University, Newcastle upon Tyne, UK)
Edmond S. L. Ho (Department of Computer and Information Sciences, Northumbria University, Newcastle upon Tyne, UK)
Hubert P. H. Shum (Department of Computer and Information Sciences, Northumbria University, Newcastle upon Tyne, UK)
Garry Elvin (Department of Computer and Information Sciences, Northumbria University, Newcastle upon Tyne, UK)

Journal of Enterprise Information Management

ISSN: 1741-0398

Article publication date: 17 August 2020

133

Abstract

Purpose

A core challenge in background subtraction (BGS) is handling videos with sudden illumination changes in consecutive frames. In our pilot study published in, Sakkos:SKIMA 2019, we tackle the problem from a data point-of-view using data augmentation. Our method performs data augmentation that not only creates endless data on the fly but also features semantic transformations of illumination which enhance the generalisation of the model.

Design/methodology/approach

In our pilot study published in SKIMA 2019, the proposed framework successfully simulates flashes and shadows by applying the Euclidean distance transform over a binary mask generated randomly. In this paper, we further enhance the data augmentation framework by proposing new variations in image appearance both locally and globally.

Findings

Experimental results demonstrate the contribution of the synthetics in the ability of the models to perform BGS even when significant illumination changes take place.

Originality/value

Such data augmentation allows us to effectively train an illumination-invariant deep learning model for BGS. We further propose a post-processing method that removes noise from the output binary map of segmentation, resulting in a cleaner, more accurate segmentation map that can generalise to multiple scenes of different conditions. We show that it is possible to train deep learning models even with very limited training samples. The source code of the project is made publicly available at https://github.com/dksakkos/illumination_augmentation

Keywords

Acknowledgements

The project was supported in part by the Royal Society (Ref: IES\R1\191147 and IES\R2\181024) and Defence and Security Accelerator (Ref: DSTLX-1000140725).

Citation

Sakkos, D., Ho, E.S.L., Shum, H.P.H. and Elvin, G. (2020), "Image editing-based data augmentation for illumination-insensitive background subtraction", Journal of Enterprise Information Management, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/JEIM-02-2020-0042

Publisher

:

Emerald Publishing Limited

Copyright © 2020, Emerald Publishing Limited

Related articles