Pith. sign in

REVIEW 12 cited by

Is synthetic data from generative models ready for image recognition?

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2210.07574 v2 pith:N5JSKTAI submitted 2022-10-14 cs.CV

Ruifei He , Shuyang Sun , Xin Yu , Chuhui Xue , Wenqing Zhang , Philip Torr , Song Bai , Xiaojuan Qi This is my paper
classification cs.CV
keywords syntheticdatamodelsrecognitionimagestasksgeneratedgeneration
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Recent text-to-image generation models have shown promising results in generating high-fidelity photo-realistic images. Though the results are astonishing to human eyes, how applicable these generated images are for recognition tasks remains under-explored. In this work, we extensively study whether and how synthetic images generated from state-of-the-art text-to-image generation models can be used for image recognition tasks, and focus on two perspectives: synthetic data for improving classification models in data-scarce settings (i.e. zero-shot and few-shot), and synthetic data for large-scale model pre-training for transfer learning. We showcase the powerfulness and shortcomings of synthetic data from existing generative models, and propose strategies for better applying synthetic data for recognition tasks. Code: https://github.com/CVMI-Lab/SyntheticData.

Discussion (0). Sign in to comment.

Forward citations

Cited by 12 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Exploring Cross-Modal Flows for Few-Shot Learning

    cs.CV 2025-10 unverdicted novelty 7.0 of 10

    FMA introduces flow matching for multi-step cross-modal feature alignment in few-shot learning, using fixed coupling, noise augmentation, and early-stopping to outperform one-step PEFT methods.

  2. An Empirical Study of Validating Synthetic Data for Text-Based Person Retrieval

    cs.CV 2025-03 unverdicted novelty 7.0 of 10

    Empirical study of a fully synthetic data generation pipeline for text-based person retrieval that tests its use as a replacement or augmentation for real data across scenarios.

  3. What Makes Synthetic Data Effective in Image Segmentation

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    Dense scene composition and instance fidelity in synthetic diffusion images drive better segmentation performance; SENSE framework exploits this to improve models on Cityscapes, COCO, and ADE20K.

  4. All in One: A Unified Synthetic Data Pipeline for Multimodal Video Understanding

    cs.CV 2026-04 unverdicted novelty 6.0 of 10

    A unified synthetic data generation pipeline produces unlimited annotated multimodal video data across multiple tasks, enabling models trained mostly on synthetic data to generalize effectively to real-world video und...

  5. R1-SyntheticVL: Is Synthetic Data from Generative Models Ready for Multimodal Large Language Model?

    cs.LG 2026-02 conditional novelty 6.0 of 10

    A collective adversarial data-synthesis pipeline produces 20K synthetic multimodal training examples whose GRPO-trained 7B model beats several listed open-source MLLMs on reasoning benchmarks.

  6. Noisy Label Refinement with Semantically Reliable Synthetic Images

    cs.CV 2025-09 conditional novelty 6.0 of 10

    Synthetic images generated per class serve as stable feature anchors; combining feature similarity to these anchors with classifier confidence corrects noisy labels and improves classification accuracy, especially und...

  7. AC3S: Adaptive Conditioning for 3D-Aware Synthetic Data Generation

    cs.CV 2026-06 unverdicted novelty 5.0 of 10

    AC3S adds a self-supervised visual prompt modulator to ControlNet diffusion and a multi-agent VLM prompt composer to generate photorealistic images with accurate 2D/3D annotations while avoiding over-conditioning.

  8. Personalized Generative Models for Contextual Debiasing

    cs.CV 2026-05 unverdicted novelty 5.0 of 10

    DecoupleGen personalizes diffusion models to create images with uncommon contexts for debiasing object recognition, yielding consistent gains on scene classification tasks.

  9. Mutual Enhancement Between Global Tokens and Patch Tokens: From Theory to Practice

    cs.CV 2026-05 unverdicted novelty 5.0 of 10

    TaTok is a theoretically grounded adaptive tokenization method that uses global tokens and cumulative conditional entropy filtering to reduce redundancy while improving reconstruction quality over fixed-rate patch tok...

  10. Scaling Up Occupancy-centric Driving Scene Generation: Dataset and Method

    cs.CV 2025-10 conditional novelty 5.0 of 10

    UniScenev2 scales occupancy-centric driving-scene generation to NuPlan scale, releasing a 3.6M-frame semantic-occupancy dataset and jointly generating occupancy, video, and LiDAR that beats published baselines on its ...

  11. DoorDet: Semi-Automated Multi-Class Door Detection Dataset via Object Detection and Large Language Models

    cs.CV 2025-08 unverdicted novelty 5.0 of 10

    The abstract and body describe different papers; the body proposes TriReWeight, a re-weighting wrapper for generative data augmentation claimed to add 2.9 to 7.9 accuracy points in small-dataset classification.

  12. Enhancing Zero-Shot Pedestrian Attribute Recognition with Synthetic Data Generation: A Comparative Study with Image-To-Image Diffusion Models

    cs.CV 2025-09 conditional novelty 4.0 of 10

    Diffusion-based synthetic expansion of pedestrian attribute training sets yields 1.1 to 3.3 point mA gains over the base PAR model on PA100k, PETAzs, and RAPzs.

Pith tools