AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Beyond Objects: Contextual Synthetic Data Generation for Fine-Grained Classification

William Yang, Xindi Wu, Zhiwei Deng, Esin Tureci, Olga Russakovsky

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Personalized and region-controlled generation

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add context preservation and context marginalization (BOB) to DataDream-style T2I fine-tuningDataDream · Aircraft 5-shotaccuracy: 50.04 -> 57.37Improved
add context marginalization (sample background/pose across dataset) without preservation (ablation)DataDream · Aircraft 10-shot, ImageNet ResNet-50accuracy: 68.00 -> 70.13Improved
add context preservation (attribute-enriched captions) without marginalization (ablation)DataDream · Aircraft 10-shot, ImageNet ResNet-50accuracy: 68.00 -> 65.90Hurt
add context preservation on top of marginalization (ablation)BOB · Aircraft 10-shot, ImageNet ResNet-50accuracy: 70.13 -> 73.78Improved
sample contexts across the whole dataset instead of only from the same class (ablation)BOB · Aircraft 10-shot, ImageNet ResNet-50accuracy: 64.38 -> 73.78Improved
use all extracted dataset contexts instead of 100 GPT-generated backgrounds/poses (ablation)BOB · Aircraft 10-shot, ImageNet ResNet-50accuracy: 72.10 -> 73.78Improved
use GPT-4o instead of Qwen2.5-VL-7B as captioning model (ablation)BOB · Aircraft 10-shot, ImageNet ResNet-50accuracyNo clear effect
use weaker Qwen2.5-VL-3B captioning model (ablation)BOB · Aircraft 10-shot, ImageNet ResNet-50accuracy: 70.88No clear effect
BOB instead of Diff-II on long-tail classificationDiff-II · CUB-LT IF=100accuracy: 56.10 -> 62.19Improved
BOB instead of Diff-II on easier long-tail benchmarkDiff-II · Flower-LT IF=50 and IF=10accuracy: 95.20 -> 95.68 (IF=50); 97.62 -> 96.13 (IF=10)No clear effect
BOB instead of DataDream on PetsDataDream · Pets 5-shotaccuracy: 92.67 -> 92.73No clear effect
augment 5 real images with BOB synthetic data versus 10 real images onlyImageNet ResNet-50 · Aircraftaccuracy: 55.48 -> 60.31Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy