Date created: | Last Updated:

Category: Project

Description: PolypDB is a large-scale, multi-center, and multi-modality dataset designed for the development of advanced AI algorithms in colonoscopy. The dataset comprises 3,934 polyp images from various imaging modalities and medical centers, offering a rich resource for research in polyp detection and segmentation. PolypDB addresses the critical need for robust and generalizable data for developing computer-aided diagnosis (CAD) systems. This dataset is publicly available and is designed to facilitate research and development in the medical AI community. Features Multi-Modality Data: Includes images from BLI, FICE, LCI, NBI, and WLI modalities. Multi-Center Data: Sourced from three medical centers in Norway, Sweden, and Vietnam. High-Quality Annotations: Verified by a team of 10 gastroenterologists with over 10 years of experience. Comprehensive Benchmarking: Includes benchmark results on eight popular segmentation methods and six standard detection methods. Publicly Accessible: The dataset is open for research and educational purposes. Download it here.