KeepOriginalAugment: Single Image-based Better Information-Preserving Data Augmentation Approach

TL;DR

KeepOriginalAugment is a novel data augmentation method that intelligently preserves salient regions within images, balancing diversity and information retention to improve model performance in computer vision tasks.

Contribution

This paper introduces KeepOriginalAugment, a new augmentation approach that incorporates salient regions into non-salient areas, addressing overfitting and domain shift issues of prior methods.

Findings

Outperforms existing augmentation techniques on CIFAR-10, CIFAR-100, and TinyImageNet.

Enhances model generalization by preserving crucial image information.

Demonstrates effectiveness across multiple strategies for salient region placement.

Abstract

Advanced image data augmentation techniques play a pivotal role in enhancing the training of models for diverse computer vision tasks. Notably, SalfMix and KeepAugment have emerged as popular strategies, showcasing their efficacy in boosting model performance. However, SalfMix reliance on duplicating salient features poses a risk of overfitting, potentially compromising the model's generalization capabilities. Conversely, KeepAugment, which selectively preserves salient regions and augments non-salient ones, introduces a domain shift that hinders the exchange of crucial contextual information, impeding overall model understanding. In response to these challenges, we introduce KeepOriginalAugment, a novel data augmentation approach. This method intelligently incorporates the most salient region within the non-salient area, allowing augmentation to be applied to either region. Striking a balance between data diversity and information preservation, KeepOriginalAugment enables models to leverage both diverse salient and non-salient regions, leading to enhanced performance. We explore three strategies for determining the placement of the salient region minimum, maximum, or random and investigate swapping perspective strategies to decide which part (salient or non-salient) undergoes augmentation. Our experimental evaluations, conducted on classification datasets such as CIFAR-10, CIFAR-100, and TinyImageNet, demonstrate the superior performance of KeepOriginalAugment compared to existing state-of-the-art techniques.