Brain MR Image Segmentation for Tumor Identification Using Hybrid of FCM Clustering and ResNet

Abstract

The current study focuses on brain tumor segmentation which is a relevant task for processing a medical image. The treatment possibilities can be increased by diagnosing the brain tumor at early stages which also helps in enhancing the survival rates. It is a hard and time-consuming task to perform the processes like brain tumor segmentation from huge number of magnetic resonance imaging (MRI) images created in laboratories manually for cancer diagnosis. Hence, there arises a need for the segmentation of brain images for tumor identification. In this paper, we attempted to address the existing problems and generate the image segmentation using fuzzy C-means (FCM) clustering method and then the classification of brain MR images using ResNet-50 model. A dataset provided by Kaggle comprising brain MR images has been used in training and testing the model. The image segmentation results are compared with the K-means clustering method using various performance metrics. The maximum accuracy of 91.18% has been observed at 350 epochs by the proposed method. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.