ADLDA: A Method to Reduce the Harm of Data Distribution Shift in Data Augmentation

TL;DR

ADLDA is a novel data augmentation method that reduces the negative effects of distribution shifts by partitioning augmented data and applying domain adaptation, leading to improved model robustness and accuracy in computer vision tasks.

Contribution

This paper introduces ADLDA, a new data augmentation technique that incorporates domain labels and adaptation to mitigate distribution shift effects in computer vision models.

Findings

Significantly improves model performance across multiple datasets.

Enhances the model's ability to recognize key features.

Effective in object recognition and image segmentation.

Abstract

This study introduces a novel data augmentation technique, ADLDA, aimed at mitigating the negative impact of data distribution shifts caused by the data augmentation process in computer vision task. ADLDA partitions augmented data into distinct subdomains and incorporates domain labels, combined with domain adaptation techniques, to optimize data representation in the model's feature space. Experimental results demonstrate that ADLDA significantly enhances model performance across multiple datasets, particularly in neural network architectures with complex feature extraction layers. Furthermore, ADLDA improves the model's ability to locate and recognize key features, showcasing potential in object recognition and image segmentation tasks. This paper's contribution provides an effective data augmentation regularization method for the field of computer vision aiding in the enhancement of robustness and accuracy in deep learning models.