-We wanted to make an applied AI project to make healthcare tasks simpler and easier.
What it does
-This app segments brain tumours/masses from MRI images.
How we built it
-This app was built using Flask and Pytorch implementing a UNet model.
Challenges we ran into
-We faced challenges in integrating the Pytorch model into the Flask webapp and creating a smooth user interface.
Accomplishments that we're proud of
-Our final app was polished and aesthetically appealing, and also featured fast, real-time inferencing.
What we learned
-We learned how to build and train a computer vision segmentation model for an applied task and integrate it into a simple webapp for deployment.
What's next for Brain Tumour Segmentation App
-We would like to improve this app using more advanced segmentation architectures such as SLSDeep and SegCaps.