Medical Imaging - Brain Tumor Segmentation

The problem MedSpot solves

MedSpot can be a boon to the cancer research and detection made easy model. Machine learning can take place of certain labour/immediate medical aid and work efficiently. Cancer has always been an unending/undetectable curse to humanity. This model would be successful in The problems it solves:

1. Brain Scan reports can be directly sent to this web application. Saves effort!
2. It can detect tumours that a layman can’t.
3. Would be available for public use.

Challenges we ran into

1. The main problem encountered was during deploying the ML model onto the frontend (Based on Php). Initially, we used flask. But later on, dropped that and found another good alternative (Using shell commands).
2. There was a major problem with the ML model as it wasn’t converging properly with the appropriate labels, but we figured a way out.
3. Another important problem was Time, if the submission deadlines were longer, our model could have been more precisely trained.

Demo Video