A Deep Learning Based Diagnostic Model Using Neuro Images (Brain Stroke Diagnosis)

Abstract

The growing demand for swift and accurate
diagnosis in clinical practice has driven
innovation in automated systems, particularly for
time-sensitive conditions like stroke. Stroke
remains a major cause of mortality and long-term
disability across the world. Early intervention is
crucial in minimizing the damage caused by
interrupted blood supply to the brain. However, the
diagnostic process involving neuroimaging
techniques like CT and MRI scans is often hindered
by delays due to manual evaluation.
This project proposes a deep learning-based
diagnostic model using Convolutional Neural
Networks (CNNs) to detect stroke-affected areas
within neuroimages. By leveraging open-source
annotated datasets such as ISLES and ATLAS, this
model automates the detection of ischemic lesions,
significantly improving diagnostic speed and
reliability. Through extensive training and testing,
the system demonstrates high performance on
metrics such as accuracy, precision, recall, and
ROC-AUC. This makes it suitable f...
The growing demand for swift and accurate
diagnosis in clinical practice has driven innovation
in automated systems, particularly for timesensitive
conditions like stroke. Stroke remains a
major cause of mortality and long-term disability
across the world. Early intervention is crucial in
minimizing the damage caused by interrupted
blood supply to the brain. However, the diagnostic
process involving neuroimaging techniques like
CT and MRI scans is often hindered by delays due
to manual evaluation.
This project proposes a deep learning-based
diagnostic model using Convolutional Neural
Networks (CNNs) to detect stroke-affected areas
within neuroimages. By leveraging open-source
annotated datasets such as ISLES and ATLAS, this
model automates the detection of ischemic lesions,
significantly improving diagnostic speed and
reliability. Through extensive training and testing,
the system demonstrates high performance on
metrics such as accuracy, precision, recall, and
ROC-AUC. This makes it suitable f...