论文标题
来自生成模型的合成数据是否准备好用于图像识别？
Is synthetic data from generative models ready for image recognition?
论文作者
论文摘要
最近的文本到图像生成模型在产生高保真照片现实图像方面显示出令人鼓舞的结果。尽管结果对人眼令人惊讶，但这些生成的图像对于识别任务的适用程度仍然不足。在这项工作中，我们广泛研究了是否以及如何将合成图像从最先进的文本到图像生成模型中用于图像识别任务，并专注于两个观点：用于改进数据筛分设置中的分类模型的合成数据（即零击和少数且很少），以及用于大规模模型的合成数据，用于传输大型模型。我们展示了现有生成模型的合成数据的强大性和缺点，并提出了更好地应用合成数据识别任务的策略。代码：https：//github.com/cvmi-lab/syntheticdata。
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.
