Regional Multi-view Learning for Cardiac Motion Analysis

dataset

posted on 2021-09-13, 18:33 authored by Esther Puyol-Antón, Bram Ruijsink, Bernhard Gerber, Mihaela Silvia Amzulescu, Hélène Langet, Mathieu De Craene, Julia Schnabela, Paolo Piro, Andrew P King

EPSRC funded project to investigate machine learning based methods to use MR and Ultrasound data for motion analysis. Data acquired from healthy volunteers with ethics and consent. This project belongs to the CDT in Medical Imaging Project of the School of Biomedical Engineering & Imaging Sciences, King's College of London. The CDT in Medical Imaging Project has multiple sub-projects, and this one is called Multimodal analysis of cardiac motion and deformation. The dataset contains MR and ultrasound healthy volunteer data acquired at St Thomas Hospital but not through the NHS. All volunteers have signed the consent form that allow release of the data. Furthermore, each dataset has been anonymized and stored in Nifti format. This data supports the article Puyol-Anton, E. et al, (2018) Regional Multi-view Learning for Cardiac Motion Analysis: Application to Identification of Dilated Cardiomyopathy Patients. IEEE Transactions on Biomedical Engineering. doi: 10.1109/TBME.2018.2865669.

Data will become openly accessible on 01/06/2018 following a 6 month embargo. Creator confirms embargo lift on 14/08/2018.

Funding

EPSRC

History

Collection method

These are the acquisition parameters: MR acquisitions were performed using a 1.5T Philips Ingenia System (Philips Healthcare, Best, The Netherlands), and the US data sets were acquired using an iE33 3D echocardiography system (Philips Medical Systems, Bothell, WA, United States) with a 1–5 MHz transthoracic matrix array transducer (xMATRIX X5.1). Full-volume acquisition mode was used in which several smaller imaging sectors acquired over multiple cardiac cycles are combined to form a large composite volume.