Orcan Alpar presents Mathematical and intelligent medical image analysis in clinical brain tumor and multiple sclerosis

Tumor detection, identification, segmentation, differentiation, grading and
classification have been the basic tasks of the radiologists still mostly
executed by eye inspection in clinical brain tumor research. Likewise, in MS
research, lesion counting, diameter and volume computation steps are crucial
for
precise progression rate estimation accomplished by neuroradiologists to reach
a
consensus, which is not always possible. Therefore, we, in our projects,
address
the clinical needs and build expert systems to automate these steps with fully
mathematical frameworks for providing solutions to real-world problems.
Combining the mathematical infrastructure with the intelligent tools, we mainly
introduce novel methods for near-zero contrast lesion identification and
segmentation by axial FLAIR and fusion of multimodal MRI images in brain tumor
research with glioma grading by fractal complexity analysis and 3D
reconstruction by morphing. In our pediatric MS research, we also focus on
segmentation of the MS lesions, automatic lesion counting with computation of
lesion diameters and coordinates for accurate volume and progression rate
estimation. We’re recently dealing with 3D reconstruction of MS lesions with
Euclidean and non-Euclidean manifolds for realistic visualization and total
volume computation.