Data to 7 Oct 2026

This dataset does not exist: training models from generated images

arXiv 1911.02888 ↗First posted 7 Nov 2019cs.CV cs.LG eess.IVcited by 65

Not on the map. Why?

Abstract (arXiv)

Current generative networks are increasingly proficient in generating high-resolution realistic images. These generative networks, especially the conditional ones, can potentially become a great tool for providing new image datasets. This naturally brings the question: Can we train a classifier only on the generated data? This potential availability of nearly unlimited amounts of training data challenges standard practices for training machine learning models, which have been crafted across the years for limited and fixed size datasets. In this work we investigate this question and its related challenges. We identify ways to improve significantly the performance over naive training on randomly generated images with regular heuristics. We propose three standalone techniques that can be applied at different stages of the pipeline, i.e., data generation, training on generated data, and deploying on real data. We evaluate our proposed approaches on a subset of the ImageNet dataset and show encouraging results compared to classifiers trained on real images.

Key facts. Cited by 65 papers in the map (data through 7 Oct 2026). First posted on arXiv 7 Nov 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 8; 2022: 4; 2023: 27; 2024: 19; 2025: 5; 2026: 2.

65 papers.

Sort:

By area:

2022Is synthetic data from generative models ready for image recognition?cited by 320
2023Effective Data Augmentation With Diffusion Modelscited by 282
2022Fake it till you make it: Learning transferable representations from synthetic ImageNet clonescited by 171
2023DiffuMask: Synthesizing Images with Pixel-level Annotations for Semantic Segmentation Using Diffusion Modelscited by 162
2021Generative Models as a Data Source for Multiview Representation Learningcited by 104
2023Synthetic Data Generation with Large Language Models for Text Classification: Potential and Limitationscited by 96
2021Learning to See by Looking at Noisecited by 92
2023Diversity is Definitely Needed: Improving Model-Agnostic Zero-shot Classification via Stable Diffusioncited by 81
2023Efficient Dataset Distillation via Minimax Diffusioncited by 81
2021GAN-Supervised Dense Visual Alignmentcited by 61
2023Diffusion Models and Semi-Supervised Learners Benefit Mutually with Few Labelscited by 37
2023Real-Fake: Effective Training Data Synthesis Through Distribution Matchingcited by 35

38 more

2023Do text-free diffusion models learn discriminative visual representations?cited by 33
2021Data-Free Knowledge Transfer: A Surveycited by 31
2023Class-Incremental Learning using Diffusion Model for Distillation and Replaycited by 31
2023Toward Understanding Generative Data Augmentationcited by 25
2024Do Generated Data Always Help Contrastive Learning?cited by 25
2022Covariance Matrix Adaptation MAP-Annealingcited by 24
2022Self-Supervised Geometry-Aware Encoder for Style-Based 3D GAN Inversioncited by 24
2023Reliability in Semantic Segmentation: Can We Use Synthetic Data?cited by 24
2023Gaussian3Diff: 3D Gaussian Diffusion for 3D Full Head Synthesis and Editingcited by 22
2023CHORUS: Learning Canonicalized 3D Human-Object Spatial Relations from Unbounded Synthesized Imagescited by 19
2021Student Surpasses Teacher: Imitation Attack for Black-Box NLP APIscited by 16
2023Image retrieval outperforms diffusion models on data augmentationcited by 14
2021GANORCON: Are Generative Models Useful for Few-shot Segmentation?cited by 13
2023AI-Generated Images as Data Source: The Dawn of Synthetic Eracited by 13
2024Advancing Fine-Grained Classification by Structure and Subject Preserving Augmentationcited by 12
2025Enhancing Vision-Language Compositional Understanding with Multimodal Synthetic Datacited by 12
2023Bridging the Gap: Enhancing the Utility of Synthetic Data via Post-Processing Techniquescited by 11
2024A Simple Background Augmentation Method for Object Detection with Diffusion Modelcited by 9
2023GenQ: Quantization in Low Data Regimes with Generative Synthetic Datacited by 8
2024Active Generation for Image Classificationcited by 8
2024Controlled Training Data Generation with Diffusion Modelscited by 8
2023Instance-Conditioned GAN Data Augmentation for Representation Learningcited by 7
2023DiffAug: Enhance Unsupervised Contrastive Learning with Domain-Knowledge-Free Diffusion-based Data Augmentationcited by 7
2021Synthetic Data for Model Selectioncited by 6
2023Free-ATM: Exploring Unsupervised Learning on Diffusion-Generated Images with Free Attention Maskscited by 6
2023Semantic Generative Augmentations for Few-Shot Countingcited by 6
2024Is Synthetic Image Useful for Transfer Learning? An Investigation into Data Generation, Volume, and Utilizationcited by 6
2021Representation Consolidation for Training Expert Studentscited by 5
2023Benchmarking and Analyzing Generative Data for Visual Recognitioncited by 5
2024Synthetic Image Learning: Preserving Performance and Preventing Membership Inference Attackscited by 5
2023Improving the Effectiveness of Deep Generative Datacited by 4
2023Harnessing Synthetic Datasets: The Role of Shape Bias in Deep Neural Network Generalizationcited by 4
2024Future-Proofing Class-Incremental Learningcited by 4
2024Transfer learning with generative models for object detection on limited datasetscited by 3
2024DIAGen: Semantically Diverse Image Augmentation with Generative Models for Few-Shot Learningcited by 3
2025Does Training on Synthetic Data Make Models Less Robust?cited by 3
2023Learning to Generate Training Datasets for Robust Semantic Segmentationcited by 2
2023SYNC-CLIP: Synthetic Data Make CLIP Generalize Better in Data-Limited Scenarioscited by 2

50 most cited of 65 All 65 (JSON)