Convolutional Neural Network Based Cerebral Edema Segmentation Using Glcm

Abstract

Computed tomography (CT) is a widely used for medical technique to diagnose
tumor and various tissue abnormalities. Rapid growth in computerized medical image
segmentation has played a vital role in scientific research. It supports physicians to take the
necessary treatments in a comfortable manner with strong decision-making. In this paper, we
have proposed a method that efficiently segment the tumor by the partitioning algorithm is
based on the Convolutional Neural Network (CNN) and the Gray Level Co-occurrence Matrix
(GLCM) feature extraction. At first, the Pre-processing is used to remove the noise from the
input and to improve the quality of an image. GLCM feature is extracted for more accurately
segment the tumors from the CT image. Segmentation and classification is done on a CNN
basis. Performance evaluation metrics is attained using parameters such as accuracy, sensitivity
specificity and precision. It reveals that the proposed CNN with GLCM performs nicely to
segment the tumors and 3.1% better compared with the state-of art methods