ResizeMix: Mixing Data with Preserved Object Information and True Labels

TL;DR

ResizeMix is a simple yet effective data augmentation technique that improves neural network performance by resizing and pasting image patches, preserving object information and outperforming existing methods without extra computational cost.

Contribution

We introduce ResizeMix, a novel data augmentation method that directly resizes and pastes image regions, addressing label misallocation and object information loss issues in previous mixing strategies.

Findings

ResizeMix outperforms CutMix and saliency-guided methods.

ResizeMix improves accuracy in image classification and detection.

It requires no additional computation compared to existing methods.

Abstract

Data augmentation is a powerful technique to increase the diversity of data, which can effectively improve the generalization ability of neural networks in image recognition tasks. Recent data mixing based augmentation strategies have achieved great success. Especially, CutMix uses a simple but effective method to improve the classifiers by randomly cropping a patch from one image and pasting it on another image. To further promote the performance of CutMix, a series of works explore to use the saliency information of the image to guide the mixing. We systematically study the importance of the saliency information for mixing data, and find that the saliency information is not so necessary for promoting the augmentation performance. Furthermore, we find that the cutting based data mixing methods carry two problems of label misallocation and object information missing, which cannot be resolved simultaneously. We propose a more effective but very easily implemented method, namely ResizeMix. We mix the data by directly resizing the source image to a small patch and paste it on another image. The obtained patch preserves more substantial object information compared with conventional cut-based methods. ResizeMix shows evident advantages over CutMix and the saliency-guided methods on both image classification and object detection tasks without additional computation cost, which even outperforms most costly search-based automatic augmentation methods.