Semantic Generative Augmentations for Few-Shot Counting
This work addresses the challenge of few-shot counting for computer vision applications, offering an incremental improvement by leveraging synthetic data generation to augment small datasets.
The paper tackled the problem of few-shot class-agnostic counting by using synthetic data generated via a double-conditioned Stable Diffusion model with prompts and density maps, and enhanced diversity through caption exchange, resulting in significant improvements in counting accuracy on FSC147 and CARPK datasets.
With the availability of powerful text-to-image diffusion models, recent works have explored the use of synthetic data to improve image classification performances. These works show that it can effectively augment or even replace real data. In this work, we investigate how synthetic data can benefit few-shot class-agnostic counting. This requires to generate images that correspond to a given input number of objects. However, text-to-image models struggle to grasp the notion of count. We propose to rely on a double conditioning of Stable Diffusion with both a prompt and a density map in order to augment a training dataset for few-shot counting. Due to the small dataset size, the fine-tuned model tends to generate images close to the training images. We propose to enhance the diversity of synthesized images by exchanging captions between images thus creating unseen configurations of object types and spatial layout. Our experiments show that our diversified generation strategy significantly improves the counting accuracy of two recent and performing few-shot counting models on FSC147 and CARPK.