Neural Network Based Classification and Diagnosis of Brain Hemorrhages

The classification and diagnosis of brain hemorrhages has work out into a greatimportance diligence in early detection of hemorrhages which reduce the deathrates. The purpose of this research was to detect brain hemorrhages and classifythem and provide the patient with correct diagnosis. A possible solution to thissocial problem is to utilize predictive techniques such as sparse componentanalysis, artificial neural networks to develop a method for detection andclassification. In this study we considered a perceptron based feed forwardneural network for early detection of hemorrhages. This paper attempts to spoton consider and talk about Computer Aided Diagnosis (CAD) that chieflynecessitated in clinical diagnosis without human act. This paper introduces aRegion Severance Algorithm (RSA) for detection and location of hemorrhagesand an algorithm for finding threshold band. In this paper different data sets (CTimages) are taken from various machines and the results obtained by applyingour algorithm and those results were compared with domain expert. Furtherresearches were challenged to originate different models in study ofhemorrhages caused by hyper tension or by existing tumor in the brain