Semantic Generative Augmentations for Few-Shot Counting - Archive ouverte HAL Access content directly
Preprints, Working Papers, ... Year : 2023

Semantic Generative Augmentations for Few-Shot Counting

Abstract

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.
Fichier principal
Vignette du fichier
main.pdf (2.56 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-04259058 , version 1 (25-10-2023)

Identifiers

Cite

Perla Doubinsky, Nicolas Audebert, Michel Crucianu, Hervé Le Borgne. Semantic Generative Augmentations for Few-Shot Counting. 2023. ⟨hal-04259058⟩
87 View
40 Download

Altmetric

Share

Gmail Facebook X LinkedIn More