Use this URL to cite or link to this record in EThOS:
Title: Joint reconstruction for multi-modality imaging with common structure
Author: Ehrhardt, M. J.
ISNI:       0000 0004 8502 035X
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2015
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Imaging is a powerful tool being used in many disciplines such as engineering, physics, biology and medicine to name a few. Recent years have seen a trend that imaging modalities have been combined to create multi-modality imaging tools where different modalities acquire complementary information. For example, in medical imaging, positron emission tomography (PET) and magnetic resonance imaging (MRI) are combined to image structure and function of the human body. Another example is spectral imaging where each channel provides information about a different wave length, e.g. information about red, green and blue (RGB). Most imaging modalities do not acquire images directly but measure a quantity from which we can reconstruct an image. These inverse problems require a priori information in order to give meaningful solutions. Assumptions are often on the smoothness of the solution but other information is sometimes available, too. Many multi-modality images show a strong inter-channel correlation as they are acquired from the same anatomy in medical imaging or the same scenery in spectral imaging. However, images from different modalities are usually reconstructed separately. In this thesis we aim to exploit this correlation using the data from all modalities, that are present in the acquisition, in a joint reconstruction process with the assumption that similar structures in all channels are more likely. We propose a framework for joint reconstruction where modalities are coupled by additional information about the solution we seek. A family of priors -- called parallel level sets -- allows us to incorporate structural a priori knowledge into the reconstruction. We analyse the parallel level set priors in several aspects including their convexity and the diffusive flow generated by their variation. Several numerical examples in RGB colour imaging and in PET-MRI illustrate the gain of joint reconstruction and in particular of the parallel level set priors.
Supervisor: Arridge, S. R. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available