Brain Tumor detection using MRI Images

Team Members:

1. Sachin Karve
2. Manovikas Ramaswamy
3. Ritwik Jadhav
4. Aswin Prasad

Abstract

Considering the medicare expenses in the developed country like US and around the world, it is no more a feasible option to casually consult doctor for first level of uneasiness. Also, in the developing countries like South Africa, Nigeria, Kenya people suffer immensely when it comes to medicare and quick attention required for certain cases. Initial negligence to scenarios like this create immense complications and might lead to major health issues down the line. There has been a lot of talk about brain tumor and the consequences of its late detection.

According to latest reports it affects around 24000 adults and 3500 children per year in the US alone. One major reason why we are losing people to brain tumor is late detection of the disease. Reports state that an early detection of tumors can save up to 13% . It is unfair to not leverage the technology of our times to develop a solution for this and hence we propose the solution to detect brain tumor in early stage using MRI scans through Artificial intelligence and computer vision.

On the other hand, in USA, Insurance companies play a vital role in the health related payment procedures. Insurance companies check the health reports of patients while registering to their company or while reimbursing the amount to the patients. Insurance companies consult the hospitals or clinics where the procedure has been carried out for the authenticity or verification of the corresponding procedure. In such cases, an application which directly gives a result to the insurance companies which verifies whether the patients has the tumor or not is highly advantageous. Insurance companies or companies can freely use such service to get an idea of the situation before or after consulting the doctor.

Taking such scenarios into account, we plan to develop an application which employes ML model and uses a brain tumor image dataset to verify the prescence of the tumor. It will be used to identify the tumorous MRI scans in real time and give out the results on our web application. This project can be extended to detect the presence of Lung cancer using CT scan images of lungs.

Design thinking

Persona/Target users

1. Insurance companies - With the health expenditure in the US increasing YoY and the never before reliance on insurance companies in health industry. It is time also to start verifying the medical conditions and do some sanity checks to verify insurance claims.

2. Patients

• Insured patients can upload this MRI as an attachment to medical record submission for insurance reimbursement, this can reduce hassle by reducing the documenation needed to file a claim.
• Patients in developing countries(Where there are scarcity of good doctors) can use this facility to identify tumors in early stage and can deicde on further treatments required.

Hill statement

Improving the access to early tumor detection mechanism to people for whom finding specialised doctors are difficult. Simplyfy the insurance claim work-flow and thus reducing hassle in both patient and insurance companies side.

Architecture Diagram

Technology stack

1. IBM Watson
2. React.js
3. Node.js
4. IBM Bluemix
5. AWS EC2