Dataset Overview

Dataset Name

CADICA Dataset

Dataset Authors

Jiménez‑Partinen, Ariadna; Molina‑Cabello, Miguel A.; Thurnhofer‑Hemsi, Karl; Palomo, Esteban; Rodríguez‑Capitán, Jorge; Molina‑Ramos, Ana I.; Jiménez‑Navarro, Manuel

Release Year

2024

Content

Contains annotated coronary angiography images, where lesion regions are labeled with bounding boxes. Original bounding boxes are given as top‑left coordinates, width, and height, later converted to YOLO format (center‑normalized coordinates and dimensions).

Usage

Intended for coronary artery disease detection, especially training deep‑learning models using the YOLO algorithm.

Structure

The dataset is divided into training, validation, and test subsets, each containing CSV files that record image paths and their ground‑truth annotations.

Processing

The DatasetGenerator class organizes the data into the directory structure required by YOLO and converts bounding‑box coordinates to YOLO format, ensuring compatibility with YOLO algorithms.

Access

The dataset can be accessed via the following link: CADICA Dataset