Automated Intracranial Hemorrhage (ICH)
Detection
Using Hybrid CNN model
_____________________________________________________________
What inspired us!
The overall incidence of spontaneous ICH worldwide is 24.6 per 100,000 person-years. Approximately half of the mortality occurs within the first 24 hours, highlighting the critical importance of early and effective treatment of the same.
The deep learning frameworks have enabled more accurate and faster detection. The radiologists could scale their diagnostic efforts with the increased CPU and GPU processing power available. Time is a matter of grave concern for medical diagnosis, and early identification of the location and type of any hemorrhage can potentially add years to the life of a patient.
Innovation in the model
Model is taking average predictions from ResNet-50 with SVM and XGBoost.
Discussion
These applications will have the flair to impact the diagnostic approaches of clinicians and health care systems and the ability for individuals to lay eyes on changes to their health in real-time.
Few description