A Hybrid Deep Learning based Classification of Brain Lesion Classification in CT Image using Convolutional Neural Networks

Abstract

In this effort, a deep learning technique for segmenting and detecting hemorrhagic lesions on brain CT images is proposed. This study intends to develop a framework for deep learning convolutional neural networks for processing CT brain images with hemorrhagic strokes and picture recognition. An adaptive median filter is used as a pre-processing step to remove noise from the input image. Following preprocessing, the picture with the noise removed is supplied into the segmentation block to be divided into numerous segments for subsequent processing. In addition, the K-means clustering technique is used in the suggested network to increase segmentation accuracy. The contrast between the hemorrhagic area and healthy brain tissue is enhanced. The findings that were acquired by employing CNN Classifier were precise. To prevail the incidence of computation is indeed slow and signals only move in one direction in feed forward setups.