Cardiac magnetic resonance image classification and retrieval based on the image acquisition technique

Abstract

Magnetic resonance imaging allows a number of imaging techniques and protocols that can be used to capture the different aspects of the cardiac function and structure. The produced amount of data is huge and its classification and/or retrieval based on its visual content are necessary for educational and training purposes. In this work, we propose a method for classification and retrieving cardiac magnetic resonance images based on the type of the acquisition technique. Preliminary results are obtained from two data sets of 3175 images acquired using five different cardiac imaging techniques. The average success rate for correctly classifying and retrieving all the images was found to be 98%.