SuperpixelGridCut, SuperpixelGridMean and SuperpixelGridMix Data Augmentation

TL;DR

This paper introduces three novel superpixel-based data augmentation methods that improve image classification performance by transforming images through dropping and fusing information, outperforming existing augmentation techniques.

Contribution

The paper presents three new superpixel grid-based augmentation methods—SuperpixelGridCut, SuperpixelGridMean, and SuperpixelGridMix—that enhance dataset diversity and model accuracy.

Findings

Significant performance improvements over baseline models.

Outperforms other data augmentation methods.

Effective across various image datasets.

Abstract

A novel approach of data augmentation based on irregular superpixel decomposition is proposed. This approach called SuperpixelGridMasks permits to extend original image datasets that are required by training stages of machine learning-related analysis architectures towards increasing their performances. Three variants named SuperpixelGridCut, SuperpixelGridMean and SuperpixelGridMix are presented. These grid-based methods produce a new style of image transformations using the dropping and fusing of information. Extensive experiments using various image classification models and datasets show that baseline performances can be significantly outperformed using our methods. The comparative study also shows that our methods can overpass the performances of other data augmentations. Experimental results obtained over image recognition datasets of varied natures show the efficiency of these new methods. SuperpixelGridCut, SuperpixelGridMean and SuperpixelGridMix codes are publicly available at https://github.com/hammoudiproject/SuperpixelGridMasks