A new framework for recovery of shape of the right ventricle from GBP SPECT images
Chunguang Cao, Timothy S. Newman · 2005
Assessment of right ventricle (RV) performance plays an important role in diagnosing cardiopulmonary diseases. Shape-based statistics describing the RV (e.g., RV volume (RVV)) are very important for developing a good comprehension of the performance of the RV. Automatic determination of the RV's location, position, and volume from medical images (e.g., CT, MRI, Echocardiography) can improve speed and consistency of diagnosis since most current procedures for obtaining such statistics rely on substantial human involvement. An RV geometric model can be very useful in guiding automated procedures, although only a few methods of describing and exploiting the RV's shape have been presented, in part due to the RV's complex shape. Most existing models imprecisely describe the RV shape. In this extended abstract, our new RV geometric model and a computational framework that exploits this new model for automatic detection and extraction of the RV from Gated Blood Pool (GBP) Single Photon Emission Computed Tomography (SPECT) datasets are presented.