Hugo J. Kuijf1, Hendrik de Leeuw1, Chris J.G. Bakker1, Koen L. Vincken1
1Image Sciences Institute, University Medical Center Utrecht, Utrecht, Netherlands
Interest in cerebral microbleeds is increasing rapidly, because they are associated with vascular disease and dementia. Detection of microbleeds on MRI is done manually, which is time-consuming and has limited reproducibility. Semi-automated detection will improve rating quality and decrease rating time. The radial symmetry transform performs excellent at this task, but has a number of false positives that need to be censored afterwards. By applying the radial symmetry transform on phase gradient magnitude images, the number of false positives can be reduced significantly. This will decrease the required rating time even further.