Web Powered CT Scan Diagnosis for Brain Hemorrhage using Deep Learning

Researchers have applied deep learning algorithms for medical image recognition and classification, producing indubitable results in medical sciences and healthcare field. The aim of this paper is to provide an exhaustive solution for revelation of brain hemorrhage within a CT scan with the help of convolutional neural networks (CNN). In the beginning stages of brain bleeding, physicians face difficulties in detection of brains that may have hemorrhage, which adds to a misdiagnosis. This challenge faced by the physicians inspires our research. A professional radiologist is almost inexistent in hospital emergency rooms in low-resource countries, and an emergency medical officer must make this immediate and vital decision. For classification of CT scans, a hybrid CNN model was used and compared with traditional architecture like AlexNet and LeNet and was shown to outperform them. Using this model, we developed a webbased AI-platform to classify CT scans of brain as hemorrhage vs non-hemorrhage.