Search of Dynamic Magnetic Resonance Images Using Active Shape Model

This paper focuses on segmenting dynamic magnetic resonance (MR) images of the human heart stored in a database. Heart MR images are dynamic, as the size and shape of a person’s heart vary in time. Active Shape Model segmentation was used to segment dynamic images of the heart. The input data is a points file of a set of random points. This method is economical since points of all the frames need not be placed in the points file for search. Segmented dynamic images of an organ such as a heart would help physicians to better understand certain medical conditions. The novelty of this approach is that it allows automatic segmentation of thousands of dynamic MR images in the database, visualization of the shape variation and retrieval of similar cases of interest from the image database.