Pattern Recognition Of Cerebral Haemorrhage (CH)From CT Images

A Cerebral Haemorrhage (CH) refers to bleeding in the brain tissues. It caused immediate severe headache thus remain unconscious. At this point, if the patients were not referred immediately to the hospital, death may become the final result. The diagnosis of CH patterns so far in Malaysia, have been performed by visual interpretation of Computed Tomography (CT) images, based on radiologist’s medical experiences. As the CH stages be divided into Hyperdense, Isodense and Hypodense area, they could detect the Hyperdense region only, as it appears in brighter tone within the CT images. The main objective of this study is to characterize the CH spectral patterns in the digital image analysis domain, i.e. Mean and Standard Deviation by mapping the CH stages from various CT sets. The technique that have been implemented in this study was the Supervised Classification with the help of medical input to determine the area of those three CH stages. Results of this study indicates that, using the procedure of hypothesis testing of those three CH region, only the Mean of Hyperdense area could be accepted with 95% significance level