Image Augmentation Preserving Object Parts Using Superpixels of Variable Granularity

D. Sun, F. Dornaika, F. Dornaika

2024

Abstract

Methods employing regional dropout data augmentation, especially those employing a cut-and-paste approach, have proven highly effective in addressing overfitting challenges arising from limited data. However, existing cutmix-based augmentation strategies face issues related to the loss of contour details and discrepancies between augmented images and their associated labels. In this study, we introduce a novel end-to-end cutmix-based data augmentation method, incorporating the blending of images with discriminative superpixels of diverse granularity. Our experiments for classification tasks reveal outstanding performance across various benchmarks and deep neural network models.

Paper Citation

in Harvard Style

Sun D. and Dornaika F. (2024). Image Augmentation Preserving Object Parts Using Superpixels of Variable Granularity. In Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP; ISBN 978-989-758-679-8, SciTePress, pages 710-717. DOI: 10.5220/0012430800003660

in Bibtex Style

@conference{visapp24,
author={D. Sun and F. Dornaika},
title={Image Augmentation Preserving Object Parts Using Superpixels of Variable Granularity},
booktitle={Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP},
year={2024},
pages={710-717},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012430800003660},
isbn={978-989-758-679-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP
TI - Image Augmentation Preserving Object Parts Using Superpixels of Variable Granularity
SN - 978-989-758-679-8
AU - Sun D.
AU - Dornaika F.
PY - 2024
SP - 710
EP - 717
DO - 10.5220/0012430800003660
PB - SciTePress