Polyp-DDPM: Diffusion-Based Semantic Polyp Synthesis for Enhanced Segmentation

TL;DR

Polyp-DDPM is a diffusion-based model that generates realistic polyp images conditioned on masks, improving segmentation accuracy and data augmentation for medical imaging, addressing data scarcity and privacy issues.

Contribution

This paper presents Polyp-DDPM, a novel diffusion model that synthesizes high-quality, diverse polyp images conditioned on masks, outperforming existing methods in image quality and segmentation performance.

Findings

Achieved a FID score of 78.47, better than previous scores above 83.79.

Attained an IoU of 0.7156, surpassing baseline synthetic and real data.

Generated datasets significantly improved polyp segmentation models.

Abstract

This study introduces Polyp-DDPM, a diffusion-based method for generating realistic images of polyps conditioned on masks, aimed at enhancing the segmentation of gastrointestinal (GI) tract polyps. Our approach addresses the challenges of data limitations, high annotation costs, and privacy concerns associated with medical images. By conditioning the diffusion model on segmentation masks-binary masks that represent abnormal areas-Polyp-DDPM outperforms state-of-the-art methods in terms of image quality (achieving a Frechet Inception Distance (FID) score of 78.47, compared to scores above 83.79) and segmentation performance (achieving an Intersection over Union (IoU) of 0.7156, versus less than 0.6694 for synthetic images from baseline models and 0.7067 for real data). Our method generates a high-quality, diverse synthetic dataset for training, thereby enhancing polyp segmentation models to be comparable with real images and offering greater data augmentation capabilities to improve segmentation models. The source code and pretrained weights for Polyp-DDPM are made publicly available at https://github.com/mobaidoctor/polyp-ddpm.