A data-centric approach to class-specific bias in image data augmentation
Preprint
Abstract
Data augmentation (DA) enhances model generalization in computer vision but may introduce biases, impacting class accuracy unevenly. Our study extends this inquiry, examining DA's class-specific bias across various datasets, including those distinct from ImageNet, through random cropping. We evaluated this phenomenon with ResNet50, EfficientNetV2S, and SWIN ViT, discovering that while residual models showed similar bias effects, Vision Transformers exhibited greater robustness or altered dynamics. This suggests a nuanced approach to model selection, emphasizing bias mitigation. We also refined a "data augmentation robustness scouting" method to manage DA-induced biases more efficiently, reducing computational demands significantly (training 112 models instead of 1860; a reduction of factor 16.2) while still capturing essential bias trends.Keywords
All Related Versions
This publication has 5 references indexed in Scilit:
- Computer vision algorithms and hardware implementations: A surveyIntegration, 2019
- A survey on Image Data Augmentation for Deep LearningJournal of Big Data, 2019
- Deep Residual Learning for Image Recognition2016
- Data Augmentation for Deep Neural Network Acoustic ModelingIEEE/ACM Transactions on Audio, Speech, and Language Processing, 2015
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998