A multilayered neural network for processing 2D tomographic images in neurosurgery

Summary form only given. A prototype artificial neural network (ANN) which is capable of processing serial sections of the brain obtained from computed-tomography (CT) or magnetic resonance imaging (MRI) tomographs has been designed. The segmented, outlined images, representing internal brain structures, both normal and abnormal, are then used as an input to a 3-D stereotaxic radiosurgery planning software. The algorithm was implemented as a software simulation in a microcomputer, structured as three cascading subnetworks. The ANN performs very well in the overall task of obtaining automatically outlined and segmented brain slices, for the purposes of 3-D reconstruction and surgical planning