An efficient computer aided detection of cerebral microbleeds from mri images

Abstract

Tiny focal points of chronic brain hemorrhages caused by blood vessels structural abnormality that makes blood product deposits are called Cerebral Microbleeds (CMBs). The occurrence of CMBs in the brain may cause traumatic injury in the brain, dementia, and amyloid angiopathy as well as lead to physiological complications with a state of confusing behavior which makes it difficult to do day-to-day activities. Within an image, the subsistence of CMB may be multiple and the number of CMBs present in the brain image determines the riskiness of the disease. The welling of the blood vessels is the main cause of CMB. Manual detection of CMB will result in errors, increased time consumption, subjectivity with limited reproducibility, and incorrect diagnosis due to its small size. Due to the small size and the erratic distribution of tiny CMBs all through the brain image, it is difficult to detect CMB automatically. The most effective diagnostic methods for identifying CMBs are Computed Tomography (CT) and Magnetic Resonance Imaging (MRI). The ionizing radiation outcome of CT scanners makes them less used in CMB detection and hence MRI images are preferred most for detecting CMB. Because CMB is more visible in the Susceptibility Weighted Imaging (SWI) type of MRI, radiologists favor SWI as suitable for the perfect identification of CMBs. Hence an effective computer-aided detection of CMB from MRI brain image is an essential need. newline