Datasets for Cardiac MRI Segmentation using BS2GDL (ACDC and LVQuan19)

This dataset contains cardiac cine-MRI images and corresponding segmentation masks used in the study entitled "BS2GDL: B-spline Geometrically-Guided Deep Learning for Left Ventricle Segmentation". The data are derived from two publicly available benchmarks: the ACDC dataset and the LVQuan19 dataset. They have been preprocessed and organized to support training and evaluation of deep learning models for left ventricle (LV) segmentation. The dataset includes:- ACDC subset: 2236 2D cine-MRI images with corresponding endocardial and epicardial masks, and a test set including 61 patients.- LVQuan19 subset: 1008 2D cine-MRI images with corresponding segmentation masks, and a test set including 7 patients. All images are aligned with their corresponding masks and follow a consistent naming convention. The data were used to train and evaluate the proposed BS2GDL model, as well as to benchmark its performance against existing methods. This repository provides the minimal dataset required to reproduce the main results presented in the associated publication, including segmentation experiments and quantitative evaluations. Note that the original datasets (ACDC and LVQuan19) are subject to their respective licenses. This repository contains processed data intended for research and reproducibility purposes. For full details on preprocessing steps, model architecture, and experimental setup, please refer to the associated publication and the accompanying code repository.