Diffusion Model with Clustering-based Conditioning for Food Image Generation

TL;DR

This paper introduces ClusDiff, a clustering-based conditional diffusion model that generates high-quality synthetic food images to improve data augmentation and class balance in food image analysis.

Contribution

The paper presents ClusDiff, a novel clustering-based training framework for diffusion models, enhancing synthetic food image quality and diversity for data augmentation.

Findings

ClusDiff outperforms existing generative models on Food-101.

Synthetic images from ClusDiff improve food classification accuracy.

ClusDiff helps mitigate class imbalance in food datasets.

Abstract

Image-based dietary assessment serves as an efficient and accurate solution for recording and analyzing nutrition intake using eating occasion images as input. Deep learning-based techniques are commonly used to perform image analysis such as food classification, segmentation, and portion size estimation, which rely on large amounts of food images with annotations for training. However, such data dependency poses significant barriers to real-world applications, because acquiring a substantial, diverse, and balanced set of food images can be challenging. One potential solution is to use synthetic food images for data augmentation. Although existing work has explored the use of generative adversarial networks (GAN) based structures for generation, the quality of synthetic food images still remains subpar. In addition, while diffusion-based generative models have shown promising results for general image generation tasks, the generation of food images can be challenging due to the substantial intra-class variance. In this paper, we investigate the generation of synthetic food images based on the conditional diffusion model and propose an effective clustering-based training framework, named ClusDiff, for generating high-quality and representative food images. The proposed method is evaluated on the Food-101 dataset and shows improved performance when compared with existing image generation works. We also demonstrate that the synthetic food images generated by ClusDiff can help address the severe class imbalance issue in long-tailed food classification using the VFN-LT dataset.