# Synthetic Examples Improve Generalization for Rare Classes

arXiv: https://arxiv.org/abs/1904.05916 · HTML: https://mlsanity.com/paper/1904.05916/

## Key facts

- Cited by 35 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 11 Apr 2019.

## Built on by (most cited first, 30 of 35)

- [A Fine-Grained Analysis on Distribution Shift](https://mlsanity.com/paper/2110.11328/) (2021)
- [Distributional Robustness Loss for Long-tail Learning](https://mlsanity.com/paper/2104.03066/) (2021)
- [The iWildCam 2021 Competition Dataset](https://mlsanity.com/paper/2105.03494/) (2021)
- [On Pitfalls of Test-Time Adaptation](https://mlsanity.com/paper/2306.03536/) (2023)
- [3D-VField: Adversarial Augmentation of Point Clouds for Domain Generalization in 3D Object Detection](https://mlsanity.com/paper/2112.04764/) (2021)
- [Perspectives on individual animal identification from biology and computer vision](https://mlsanity.com/paper/2103.00560/) (2021)
- [Out-of-Domain Robustness via Targeted Augmentations](https://mlsanity.com/paper/2302.11861/) (2023)
- [Simulated Adversarial Testing of Face Recognition Models](https://mlsanity.com/paper/2106.04569/) (2021)
- [Segmenting Known Objects and Unseen Unknowns without Prior Knowledge](https://mlsanity.com/paper/2209.05407/) (2022)
- [A Holistic Assessment of the Reliability of Machine Learning Systems](https://mlsanity.com/paper/2307.10586/) (2023)
- [ElephantBook: A Semi-Automated Human-in-the-Loop System for Elephant Re-Identification](https://mlsanity.com/paper/2106.15083/) (2021)
- [Uncertainty-guided Boundary Learning for Imbalanced Social Event Detection](https://mlsanity.com/paper/2310.19247/) (2023)
- [Choosing an Appropriate Platform and Workflow for Processing Camera Trap Data using Artificial Intelligence](https://mlsanity.com/paper/2202.02283/) (2022)
- [Finding Differences Between Transformers and ConvNets Using Counterfactual Simulation Testing](https://mlsanity.com/paper/2211.16499/) (2022)
- [Diffusion Curriculum: Synthetic-to-Real Data Curriculum via Image-Guided Diffusion](https://mlsanity.com/paper/2410.13674/) (2024)
- [SIMPL: Generating Synthetic Overhead Imagery to Address Zero-shot and Few-Shot Detection Problems](https://mlsanity.com/paper/2106.15681/) (2021)
- [3D Adversarial Augmentations for Robust Out-of-Domain Predictions](https://mlsanity.com/paper/2308.15479/) (2023)
- [Domain Adaptation for Rare Classes Augmented with Synthetic Samples](https://mlsanity.com/paper/2110.12216/) (2021)
- [Human Body Measurement Estimation with Adversarial Augmentation](https://mlsanity.com/paper/2210.05667/) (2022)
- [A Shared Representation for Photorealistic Driving Simulators](https://mlsanity.com/paper/2112.05134/) (2021)
- [Weighted Risk Invariance: Domain Generalization under Invariant Feature Shift](https://mlsanity.com/paper/2407.18428/) (2024)
- [ZebraPose: Zebra Detection and Pose Estimation using only Synthetic Data](https://mlsanity.com/paper/2408.10831/) (2024)
- [Lacking Data? No worries! How synthetic images can alleviate image scarcity in wildlife surveys: a case study with muskox (Ovibos moschatus)](https://mlsanity.com/paper/2511.11882/) (2025)
- [Image-to-Image Translation of Synthetic Samples for Rare Classes](https://mlsanity.com/paper/2106.12212/) (2021)
- [Elephants and Algorithms: A Review of the Current and Future Role of AI in Elephant Monitoring](https://mlsanity.com/paper/2306.13803/) (2023)
- [Examining Pathological Bias in a Generative Adversarial Network Discriminator: A Case Study on a StyleGAN3 Model](https://mlsanity.com/paper/2402.09786/) (2024)
- [Adapting the re-ID challenge for static sensors](https://mlsanity.com/paper/2412.00290/) (2024)
- [Exploring the Equivalence of Closed-Set Generative and Real Data Augmentation in Image Classification](https://mlsanity.com/paper/2508.09550/) (2025)
- [Degradation-based augmented training for robust individual animal re-identification](https://mlsanity.com/paper/2603.04163/) (2026)
- [Generating Synthetic Wildlife Health Data from Camera Trap Imagery: A Pipeline for Alopecia and Body Condition Training Data](https://mlsanity.com/paper/2603.26754/) (2026)

## Abstract (arXiv)

The ability to detect and classify rare occurrences in images has important applications - for example, counting rare and endangered species when studying biodiversity, or detecting infrequent traffic scenarios that pose a danger to self-driving cars. Few-shot learning is an open problem: current computer vision systems struggle to categorize objects they have seen only rarely during training, and collecting a sufficient number of training examples of rare events is often challenging and expensive, and sometimes outright impossible. We explore in depth an approach to this problem: complementing the few available training images with ad-hoc simulated data. Our testbed is animal species classification, which has a real-world long-tailed distribution. We analyze the effect of different axes of variation in simulation, such as pose, lighting, model, and simulation method, and we prescribe best practices for efficiently incorporating simulated data for real-world performance gain. Our experiments reveal that synthetic data can considerably reduce error rates for classes that are rare, that as the amount of simulated data is increased, accuracy on the target class improves, and that high variation of simulated data provides maximum performance gain.