BibTeX entry

@PHDTHESIS{200112Albert_Chi_Shing_Chung,
  AUTHOR={(Albert) Chi Shing Chung},
  TITLE={Vessel and Aneurysm Reconstruction using Speed and Flow Coherence
    Information in Phase Contrast Magnetic Resonance Angiograms},
  SCHOOL={University of Oxford},
  YEAR=2001,
  URL={http://www.bmva.org/theses/2001/2001-chung.pdf},
}

Abstract

Phase contrast magnetic resonance angiography (PC-MRA) is a non-invasive method for 3D vessel delineation, which for each voxel not only provides measurement of speed (conveyed as a speed image), but also gives a three-component estimate of flow direction (in the form of phase images). In this thesis, we present a new approach to reconstructing vessels and aneurysms from PC-MRA, and demonstrate how speed and flow coherence information extracted from a PC-MRA dataset can be combined for detecting and reconstructing normal vessels and aneurysms with relatively low flow rate and low signal-to-noise ratio (SNR).

We propose to use a Maxwell-Gaussian mixture density to model the background signal and combine this with a uniform distribution for modelling vascular signal to give a MaxwellGaussian-uniform (MGU) mixture model of speed image intensity. The MGU model parameters are estimated by the Expectation-Maximisation (EM) algorithm. It is shown that the Maxwell-Gaussian mixture distribution models the background signal more accurately than a Maxwell distribution. Although the MGU model works satisfactorily in classifying the background and vessel voxels, for relatively low flow rate and low SNR vessel regions (especially inside an aneurysm), we find that it is hard to distinguish vessel voxels from the background voxels because of their low intensity value.

To deal with this problem, we propose to include the information about local flow coherence as a priori knowledge modelled by a Markov random field (MRF). A new coherence measure, namely local phase coherence, which incorporates information about the spatial relationships between neighbouring flow vectors, is defined and shown to be more robust to noise than prior coherence measures. The MGU statistical measure from the speed images and the local phase coherence measure from the phase images are combined in a Bayesian framework to estimate the posterior probabilities of vessel and background. It is shown that segmentation based on speed and flow coherence information gives a higher accuracy at low and high SNR values than segmentation using either speed or flow coherence information alone. Finally, the vessels and aneurysms are reconstructed by using a sub-voxel based level set method running on the estimated posterior probabilistic maps of vessel and background.

The new method is tested on an aneurysm phantom data set and 7 clinical data sets. The results show that the proposed method can help detecting and reconstructing relatively low flow and low SNR regions when both speed and flow coherence information are utilised.