Intracranial Hemorrhage Detection on CT Scan Images using Transfer Learning Approach of Convolutional Neural Network

Abstract

Listen

Translate article

Intracranial hemorrhage is an acute bleeding within the skull when a brain’s blood vessel is ruptured which eventually leads to disability or even death. It is extremely important to diagnose intracranial hemorrhage with high accuracy to minimize the severity of brain damage and other complications. Deep Learning is widely used in interpreting medical images and has shown promising advancements in diagnosing brain hemorrhage. This paper proposes a deep learning method called Convolutional Neural Network (CNN) on neuroimaging with transfer learning techniques to assist in the diagnosis of intracranial hemorrhage on CT scans. We have used six pre-trained CNN models (EfficientNetB6, DenseNet121, ResNet50, InceptionResNetV2, InceptionV3, VGG16) and also present a traditional CNN model of 11-layer architecture for the detection and binary classification of intracranial brain hemorrhage on CT scans. The paper depicts a comparative analysis on the performance between the proposed traditional and pre-trained models of CNN in terms of accuracy, precision, recall, F1 score, and AUC curve. EfficientNetB6 yields an accuracy of 95.99%, which is higher than any of the experimental results of the CNN models used in this dataset. Lastly, we deploy a simple web application to demonstrate real-world application.