# Evaluation of Randomization through Style Transfer for Enhanced Domain Generalization

arXiv: https://arxiv.org/abs/2604.05616 · HTML: https://mlsanity.com/paper/2604.05616/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 32 references that are in the map.
- First posted on arXiv 7 Apr 2026.

## Builds on (most cited first, 24 of 24)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [The Cityscapes Dataset for Semantic Urban Scene Understanding](https://mlsanity.com/paper/1604.01685/) (2016)
- [DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs](https://mlsanity.com/paper/1606.00915/) (2016)
- [Rethinking Atrous Convolution for Semantic Image Segmentation](https://mlsanity.com/paper/1706.05587/) (2017)
- [Arbitrary Style Transfer in Real-time with Adaptive Instance Normalization](https://mlsanity.com/paper/1703.06868/) (2017)
- [BDD100K: A Diverse Driving Dataset for Heterogeneous Multitask Learning](https://mlsanity.com/paper/1805.04687/) (2018)
- [Unified Perceptual Parsing for Scene Understanding](https://mlsanity.com/paper/1807.10221/) (2018)
- [Playing for Data: Ground Truth from Computer Games](https://mlsanity.com/paper/1608.02192/) (2016)
- [Sim-to-Real Transfer in Deep Reinforcement Learning for Robotics: a Survey](https://mlsanity.com/paper/2009.13303/) (2020)
- [DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation](https://mlsanity.com/paper/2111.14887/) (2021)
- [Domain Randomization and Pyramid Consistency: Simulation-to-Real Generalization without Accessing Target Domain Data](https://mlsanity.com/paper/1909.00889/) (2019)
- [Neural Style Transfer: A Review](https://mlsanity.com/paper/1705.04058/) (2017)
- [Semantic-Aware Domain Generalized Segmentation](https://mlsanity.com/paper/2204.00822/) (2022)
- [WildNet: Learning Domain Generalized Semantic Segmentation from the Wild](https://mlsanity.com/paper/2204.01446/) (2022)
- [Global and Local Texture Randomization for Synthetic-to-Real Semantic Segmentation](https://mlsanity.com/paper/2108.02376/) (2021)
- [Frustratingly Simple Domain Generalization via Image Stylization](https://mlsanity.com/paper/2006.11207/) (2020)
- [Generalization by Adaptation: Diffusion-Based Domain Extension for Domain-Generalized Semantic Segmentation](https://mlsanity.com/paper/2312.01850/) (2023)
- [Rethinking Domain Generalization Baselines](https://mlsanity.com/paper/2101.09060/) (2021)
- [WEDGE: Web-Image Assisted Domain Generalization for Semantic Segmentation](https://mlsanity.com/paper/2109.14196/) (2021)
- [STaDA: Style Transfer as Data Augmentation](https://mlsanity.com/paper/1909.01056/) (2019)
- [Intra-Source Style Augmentation for Improved Domain Generalization](https://mlsanity.com/paper/2210.10175/) (2022)
- [Back-to-Bones: Rediscovering the Role of Backbones in Domain Generalization](https://mlsanity.com/paper/2209.01121/) (2022)
- [Data Augmentation Through Random Style Replacement](https://mlsanity.com/paper/2504.10563/) (2025)

## Abstract (arXiv)

Deep learning models for computer vision often suffer from poor generalization when deployed in real-world settings, especially when trained on synthetic data due to the well-known Sim2Real gap. Despite the growing popularity of style transfer as a data augmentation strategy for domain generalization, the literature contains unresolved contradictions regarding three key design axes: the diversity of the style pool, the role of texture complexity, and the choice of style source. We present a systematic empirical study that isolates and evaluates each of these factors for driving scene understanding, resolving inconsistencies in prior work. Our findings show that (i) expanding the style pool yields larger gains than repeated augmentation with few styles, (ii) texture complexity has no significant effect when the pool is sufficiently large, and (iii) diverse artistic styles outperform domain-aligned alternatives. Guided by these insights, we derive StyleMixDG (Style-Mixing for Domain Generalization), a lightweight, model-agnostic augmentation recipe that requires no architectural modifications or additional losses. Evaluated on the GTAV rightarrow BDD100k, Cityscapes, Mapillary Vistas benchmark, StyleMixDG demonstrates consistent improvements over strong baselines, confirming that the empirically identified design principles translate into practical gains. The code will be released on GitHub.