Medical Image Computing and Signal Processing Lab

Since Magnetic Resonance Imaging (MRI) is widely used radiological tool, need of comprehensive and quantitative evaluation of flow imaging is important. MR angiograms provide flow information in all three directions. The need for analyzing blood flow is geared towards the advancement of MRA to Phase Contrast MRA (PC-MRA) in which the vessel structures are enhanced and background tissues are suppressed using statistical post-processing.

The basis of PC-MRA is that the phase shift developed in the spins moving in the direction of a magnetic field gradient is proportional to their velocity. Though, PC-MRA provides static phase suppression, it requires pre-processing as well as post-processing for enhancement of vessel structures. The parameters provided at the time of data acquisition plays a key role in controlling the image quality. Flow modeling provides wide scope to understand PC-MRA Simulation. The statistical analysis of PC-MRA provides the need for implementing segmentation for enhancing the vascular structure. Application of Finite Mixture Model (FMM) provides scope for differentiating vascular structure from background tissue.

In this work, a study of MRI Simulation is undertaken and extended to PC-MRA Simulation. Finite Mixture Model (FMM) estimation is performed with the Expectation Maximization (EM) algorithm. The segmentation is performed using local phase coherence (LPC) by using Markov Random Field (MRF) formulation.