Multi-modal Medical Data Registration

From Master Projects
Jump to: navigation, search


has title::Multi-modal Medical Data Registration
status: finished
Master: project within::Multimedia
Student name: student name::Andra Maria Pascale
Dates
Start start date:=2012/01/01
End end date:=2012/08/31
Supervision
Supervisor: Thilo Kielmann
Second supervisor: Arjen van Rhijn
Second reader: has second reader::Ben van Werkhoven
Company: has company::Personal Space Technologies
Thesis: has thesis::Media:Thesis.pdf
Poster: has poster::Media:Posternaam.pdf

Signature supervisor



..................................

Abstract

Multimodal datasets contain physical properties obtained with different imaging modalities. In medicine, four-dimensional data (e.g. spatial-temporal) are widely used, five-dimensional data (e.g. spatialtemporal- spectral) are emerging, and multimodal data are being used more often every day. The goal of medical image registration is to geometrically align two or more image volumes or surfaces so that voxels representing the same anatomical structure may be superimposed. Modern imaging modalities may for instance provide structural, functional or metabolic information, and the combination of such information is often vital for accurate diagnosis and evaluation of treatment. The Personal Space Station (PSS) is an interactive desktop Virtual Reality system developed by PS-tech. It enables a user to view and analyze data directly in 3D and provides co-located intuitive 3D interaction. Due to its high resolution stereoscopic visualization and direct interaction, it provides several opportunities which can be exploited to simplify the image registration problem. For instance, 3D interaction techniques could be used to let a user define a rough initial registration to align two data sets, followed by an automated procedure to find a more optimal registration. This exploits the fact that registration algorithms are generally quite effective in optimization problems with a reasonable initial estimate, while human beings are generally more efficient in obtaining such an initial estimate. The work of this research has the following expected outcome: A system to perform semi-automated 3D data alignment in the PSS. Alignment will be done by automatic alignment algorithms, possibly exploiting user interaction to steer the computation (for instance by manual definition of certain anatomical landmarks).