INTRACRANIAL HEMORRAHAGE DETECTION USING CONVOLUTIONAL NEURAL NETWORK

Authors

  • M. Raja Suguna , A. Kalaivani

Abstract

IntracranialHemorrhage(ICH), also referred as bleeding inside the Cranium(skull), which
requires intense and rapid medical treatment. It is caused by Physical trauma and non-physical trauma such
as haemorrhagestroke, disorder with blood clotting and Anticoagulant therapy. Haemorrhage doesn’t spread
wide area in brain, it is focal brain injury, affecting particular localized area in the brain. Based on the
location, haemorrhage is classified as Intra-axial and extra axial. Intra-axial haemorrhage is bleeding inside
the brain further classified as intraparenchymal haemorrhage, intraventricular haemorrhage. Extra-axial
Haemorrhage is bleeding outside the brain having subtypes Epidural haemorrhage,Subdural haemorrhage,
Subarachnoid haemorrhage. Haemorrhage are detected by the Physicians using Computed Tomography
(CT)and MRI Scans. Density of blood is higher than that of brain tissue but lesser than bone,henceICH
require more precise medical detection. In most of the Emergencycentre, Patients head CT is initially
interpreted by radiologist trainee or junior radiologist and further reviewed by experienced and senior
radiologist,which leads to the discrepancies and may results in severe medical consequences.Therefore,
AutomatedSystem is necessary for the accurate detection of ICH.Recent Advances in Artificial Intelligence
in medical Image processing using deep learning algorithm gives promising result.In this paper,we proposed
a Two-Dimensionalconvolutional neural network with Long Short -Term Memory units for the accurate
detection of Intracranial Hemorrhage and its subtypes.

Published

2020-01-31

Issue

Section

Articles