Experimental Approach for Brain Region Segmentation using CNN

Artificial neural networks or CNNs are often used to solve a range of CV2 / Computer vision problems. Image classification has recently advanced to the point that it can be used for disease diagnosis. Although most Convolutional neural network utilize 2D kernels, 3D kernels were recently published in Convolutional neural network papers on clinical image analysis, enabling complete control to the 3D structures from clinical data for disease diagnosis. Clinical picture classification, while strongly tied to feature extraction, has unique problems, including the lack of labeled images, the significant group asymmetry observed in the real data, and the huge computational power demand of 3D data like points clouds for voxelization tasks. This paper shows how to use a Convolutional neural network utilizing approach with 3D filtration on brain CT scans images or Magnetic resonance imaging for Brain Region segmentation. Multiple changes to an existing Convolutional neural network design are addressed, as well as techniques for dealing with the problems presented. Whereas the majority of the previous research on clinical procedures to follow focuses on vital organ and connective tissues, our approach has been verified using datasets from both the CNS, CT scans and magnetic resonance imaging.