Diversify, Don't Fine-Tune: Scaling Up Visual Recognition Training with Synthetic Images

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Yu, Zhuoran, Zhu, Chenchen, Culatana, Sean, Krishnamoorthi, Raghuraman, Xiao, Fanyi, Lee, Yong Jae
Formato: Preprint
Publicado: 2023
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910790626312192
author Yu, Zhuoran
Zhu, Chenchen
Culatana, Sean
Krishnamoorthi, Raghuraman
Xiao, Fanyi
Lee, Yong Jae
author_facet Yu, Zhuoran
Zhu, Chenchen
Culatana, Sean
Krishnamoorthi, Raghuraman
Xiao, Fanyi
Lee, Yong Jae
contents Recent advances in generative deep learning have enabled the creation of high-quality synthetic images in text-to-image generation. Prior work shows that fine-tuning a pretrained diffusion model on ImageNet and generating synthetic training images from the finetuned model can enhance an ImageNet classifier's performance. However, performance degrades as synthetic images outnumber real ones. In this paper, we explore whether generative fine-tuning is essential for this improvement and whether it is possible to further scale up training using more synthetic data. We present a new framework leveraging off-the-shelf generative models to generate synthetic training images, addressing multiple challenges: class name ambiguity, lack of diversity in naive prompts, and domain shifts. Specifically, we leverage large language models (LLMs) and CLIP to resolve class name ambiguity. To diversify images, we propose contextualized diversification (CD) and stylized diversification (SD) methods, also prompted by LLMs. Finally, to mitigate domain shifts, we leverage domain adaptation techniques with auxiliary batch normalization for synthetic images. Our framework consistently enhances recognition model performance with more synthetic data, up to 6x of original ImageNet size showcasing the potential of synthetic data for improved recognition models and strong out-of-domain generalization.
format Preprint
id arxiv_https___arxiv_org_abs_2312_02253
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Diversify, Don't Fine-Tune: Scaling Up Visual Recognition Training with Synthetic Images
Yu, Zhuoran
Zhu, Chenchen
Culatana, Sean
Krishnamoorthi, Raghuraman
Xiao, Fanyi
Lee, Yong Jae
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Recent advances in generative deep learning have enabled the creation of high-quality synthetic images in text-to-image generation. Prior work shows that fine-tuning a pretrained diffusion model on ImageNet and generating synthetic training images from the finetuned model can enhance an ImageNet classifier's performance. However, performance degrades as synthetic images outnumber real ones. In this paper, we explore whether generative fine-tuning is essential for this improvement and whether it is possible to further scale up training using more synthetic data. We present a new framework leveraging off-the-shelf generative models to generate synthetic training images, addressing multiple challenges: class name ambiguity, lack of diversity in naive prompts, and domain shifts. Specifically, we leverage large language models (LLMs) and CLIP to resolve class name ambiguity. To diversify images, we propose contextualized diversification (CD) and stylized diversification (SD) methods, also prompted by LLMs. Finally, to mitigate domain shifts, we leverage domain adaptation techniques with auxiliary batch normalization for synthetic images. Our framework consistently enhances recognition model performance with more synthetic data, up to 6x of original ImageNet size showcasing the potential of synthetic data for improved recognition models and strong out-of-domain generalization.
title Diversify, Don't Fine-Tune: Scaling Up Visual Recognition Training with Synthetic Images
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2312.02253