Skip to content

Cross platform app which detects and classifies brain tumors using deep learning.

License

Notifications You must be signed in to change notification settings

IinjyI/Verstand

Repository files navigation

Verstand

A cross platform application built with flutter

Tech stack

Python fast api Dart firebase Flutter TensorFlow

Problem description

  • Our brain is the most complex organ of our body and therefore, the most important. However, Brain tumors are hard for humans to diagnose using MRI images only, What if we use something more accurate... more capable of finding patterns?

Goals

  • Make brain tumors diagnosis more accurate.
  • Provide means for people who can't find doctors to get a diagnosis.

About brain tumors

  • There are three types of primary brain tumors, about which we are concerned throughout this project:

    • Glioma

      • Arise from the glial cells which surround the nerve cells in the brain
      • They are the most numerous cells in the central nervous system
    • Meningioma

      • The most common type of primary tumor
      • Arises from the meninges = the lining of the brain
    • Pituitary Adenoma

      • Scientists aren’t sure of the exact cause of pituitary adenomas but some adenomas have been linked to accidental changes, or mutations in DNA

Pre-release

Project demo

Side-notes

  • The code for DL model and API are uploaded to github (private repos)

  • Due to hosting issues (I'm still working on them), if no images are sent to the API for more than 15 minutes, it becomes idle and may take up to 2-3 minutes to come back online and respond

  • MRI images has to be in .JPG/.JPEG format to be diagnosed correctly (I'm still working on this issue)