Abstract:Unsupervised Domain Adaptation (UDA) and domain generalization (DG) are two research areas that aim to tackle the lack of generalization of Deep Neural Networks (DNNs) towards unseen domains. While UDA methods have access to unlabeled target images, domain generalization does not involve any target data and only learns generalized features from a source domain. Image-style randomization or augmentation is a popular approach to improve network generalization without access to the target domain. Complex methods are often proposed that disregard the potential of simple image augmentations for out-of-domain generalization. For this reason, we systematically study the in- and out-of-domain generalization capabilities of simple, rule-based image augmentations like blur, noise, color jitter and many more. Based on a full factorial design of experiment design we provide a systematic statistical evaluation of augmentations and their interactions. Our analysis provides both, expected and unexpected, outcomes. Expected, because our experiments confirm the common scientific standard that combination of multiple different augmentations out-performs single augmentations. Unexpected, because combined augmentations perform competitive to state-of-the-art domain generalization approaches, while being significantly simpler and without training overhead. On the challenging synthetic-to-real domain shift between Synthia and Cityscapes we reach 39.5% mIoU compared to 40.9% mIoU of the best previous work. When additionally employing the recent vision transformer architecture DAFormer we outperform these benchmarks with a performance of 44.2% mIoU

Untapped Potential of Data Augmentation: A Domain Generalization Viewpoint

Boosting Unsupervised Contrastive Learning Using Diffusion-Based Data Augmentation from Scratch

Data Augmentation Revisited: Rethinking the Distribution Gap between Clean and Augmented Data

A Simple Feature Augmentation for Domain Generalization

Incorporating Supervised Domain Generalization into Data Augmentation

Selecting Data Augmentation for Simulating Interventions

DualAug: Exploiting Additional Heavy Augmentation with OOD Data Rejection

A Comprehensive Survey on Data Augmentation

Generalization Gap in Data Augmentation: Insights from Illumination

A Kernel Theory of Modern Data Augmentation

Learning to Augment via Implicit Differentiation for Domain Generalization

Data Augmentation using Generative-AI

On the Generalization Effects of Linear Transformations in Data Augmentation

Effective Data Augmentation With Diffusion Models

NormAUG: Normalization-guided Augmentation for Domain Generalization

Out-of-Domain Robustness via Targeted Augmentations

Boosting Model Resilience via Implicit Adversarial Data Augmentation

A Survey on Data Augmentation in Large Model Era

Augmentation-based Domain Generalization for Semantic Segmentation

Cross-Domain Feature Augmentation for Domain Generalization