Predavanje - assoc. prof. Gozde Unal

datum začetka: 08.08.2016
ura: 13:00
lokacija: predavalnica LST
organizator: UL FE



Univerza v Ljubljani, Fakulteta za elektrotehniko

Katedra za biomedicinsko tehniko

Vas vljudno vabi na predavanje

z naslovom:

Multivalued Data Processing Techniques for Diffusion MRI

Predaval bo:

assoc. prof. Gozde Unal

Istanbul Technical University,

Computer Engineering Department, Istanbul, Turkey

Predavanje bo:

v ponedeljek, 8. avgusta 2016, ob 13.00, v predavalnici LST (stavba A, 4. nad.)


Vljudno vabljeni!




Prof. Unal will present two novel state-of-the-art techniques for processing of MRI data in order to model and extract structural asymmetries that exist in cerebellar vessels and brain fibers. The first technique involves higher order tensors for modelling of tree-like structures such as vascular trees in the human brain. The underlying approach is to embed the tensor in a 4D space rather than 3D in order to untangle the bifurcating (or even n-furcating) structures/branches in the data in a higherdimensional space. The method is demonstrated on cerebral artery segmentation on MRA datasets and provides a seamless modeling of both the tubular and the bifurcating sections of a vascular tree within the same model. The second technique consists of an effective directional regularization approach for capturing inherent asymmetry of the underlying intravoxel geometry that exists in bending, crossing or kissing fibers of the brain white matter from Diffusion MRI data. Their study demonstrates the asymmetry of fiber microstructure at the voxel level.


Življenjepis predavatelja:

Prof. Gozde Unal obtained her PhD at North Carolina State University in 2002. She worked as a research scientist at Siemens Corporate Research, Princeton NJ, USA for four years. Later she was an Associate Professor at Sabanci University and currently holds a position of Associate Professor at Istanbul Technical University, Computer Engineering Department, Istanbul, Turkey. Her general research interests are computer vision and medical image analysis. Current research is mainly focused on segmentation, registration, and shape analysis techniques with applications to clinically relevant problems in various medical imaging modalities such as MR, US, CT, and intravascular images.