Main Article Content
The diagnosis of brain abnormalities, prognosis monitoring, and treatment evaluation all rely heavily on the magnetic resonance scan of brain tissue segmentation. Although numerous automated or semi-automatic methods have been proposed in the literature to reduce the need for human intervention, the degree of accuracy is frequently still significantly lower than that of manual segmentation. We give a clever technique for fragmenting the cerebrum utilizing a managed counterfeit brain organization system called artificial neural network (ANN) and volumetric shape models. In the beginning, in addition to the usual spatial-based and intensity-based image features, a level-set oriented brain boundary fitting technique is used to accomplish this. This is controlled by the picture intensity. The ANN is then informed of the number of important structures. Additionally, rather than directly applying standard guidelines to local appearances, this ANN learns local adaptive feature classification conditions. The outcomes demonstrate that the proposed strategy achieves competitive results in a relatively shorter time spent training.