Abstract #2154

Vascular Territory Segmentation Using Mutual Clustering in Image and Label Space

Youngkyoo Jung¹,², Megan E. Johnston², Christopher T. Whitlow¹,³, Joseph A. Maldjian¹

¹Radiology, Wake Forest School of Medicine, Winston-Salem, NC, United States; ²Biomedical Engineering, Wake Forest School of Medicine, Winston-Salem, NC, United States; ³Translational Science Institute, Wake Forest School of Medicine, Winston-Salem, NC, United States

Pseudo-continuous ASL (PCASL) based vascular territory mapping enables the detection of source arteries. An automated vascular territory segmentation using connectivity information from both image space and label space is proposed. The algorithm was tested with data obtained at two locations with middle cerebral artery branching. The algorithm demonstrates the ability to resolve neighboring sources into separate territories with minimal computational burden.