RSMDA: Random Slices Mixing Data Augmentation
-
Published:2023-01-29
Issue:3
Volume:13
Page:1711
-
ISSN:2076-3417
-
Container-title:Applied Sciences
-
language:en
-
Short-container-title:Applied Sciences
Author:
Kumar Teerath1, Mileo Alessandra2, Brennan Rob3ORCID, Bendechache Malika4ORCID
Affiliation:
1. ADAPT—Science Foundation Ireland Research Centre and CRT AI, School of Computing, Dublin City University, D02 PN40 Dublin, Ireland 2. INSIGHT Centre for Data Analytics and the I-Form Centre for Advanced Manufacturing, School of Computing, Dublin City University, D02 PN40 Dublin, Ireland 3. ADAPT, School of Computer Science, University College Dublin, D02 PN40 Dublin, Ireland 4. ADAPT & Lero Research Centres, School of Computer Science, University of Galway, H91 TK33 Galway, Ireland
Abstract
Advanced data augmentation techniques have demonstrated great success in deep learning algorithms. Among these techniques, single-image-based data augmentation (SIBDA), in which a single image’s regions are randomly erased in different ways, has shown promising results. However, randomly erasing image regions in SIBDA can cause a loss of the key discriminating features, consequently misleading neural networks and lowering their performance. To alleviate this issue, in this paper, we propose the random slices mixing data augmentation (RSMDA) technique, in which slices of one image are placed onto another image to create a third image that enriches the diversity of the data. RSMDA also mixes the labels of the original images to create an augmented label for the new image to exploit label smoothing. Furthermore, we propose and investigate three strategies for RSMDA: (i) the vertical slices mixing strategy, (ii) the horizontal slices mixing strategy, and (iii) a random mix of both strategies. Of these strategies, the horizontal slice mixing strategy shows the best performance. To validate the proposed technique, we perform several experiments using different neural networks across four datasets: fashion-MNIST, CIFAR10, CIFAR100, and STL10. The experimental results of the image classification with RSMDA showed better accuracy and robustness than the state-of-the-art (SOTA) single-image-based and multi-image-based methods. Finally, class activation maps are employed to visualize the focus of the neural network and compare maps using the SOTA data augmentation methods.
Funder
Science Foundation Ireland Lero
Subject
Fluid Flow and Transfer Processes,Computer Science Applications,Process Chemistry and Technology,General Engineering,Instrumentation,General Materials Science
Reference60 articles.
1. Novel Algorithm for Image Classification Using Cross Deep Learning Technique;Kumar;IOP Conf. Ser. Mater. Sci. Eng.,2021 2. Image classification algorithm based on deep learning-kernel function;Liu;Sci. Program.,2020 3. Research on power equipment recognition method based on image processing;Wang;EURASIP J. Image Video Process.,2019 4. Forged Character Detection Datasets: Passports, Driving Licences And Visa Stickers;Kumar;Int. J. Artif. Intell. Appl. (IJAIA),2022 5. Ciresan, D., Meier, U., Masci, J., Gambardella, L., and Schmidhuber, J. (2011, January 16–22). Flexible, high performance convolutional neural networks for image classification. Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence, Catalonia, Spain.
|
|