This is a repo for the Tanzania AI lab hackathon 2020 & the AI4Dev2020 challenge, where we as the Elixir team created the 1st AI based cancer diagnosis system, built a model comprising of Deep Convolutional Neural Network(CNN) and a web app that screens microscopic images so as to detect cancer tumors, thus increasing speed, accuracy in cancer diagnosis, and testing.
Pitch Video: https://youtu.be/Pfgfh4DJToo
This is the repository for a Cancer Diagnosis web app system, that detects cancer from microscopic images, developed during the Tanzania AI-LAB hackathon 2020 and The AI4DEV2020 Challenge
You can learn more about the app and the reason for it's existence via the following link https://www.youtube.com/watch?v=Pfgfh4DJToo
To use the web app, you need to do the following:
- Create a python virtual environment and install the packages in the requirements.txt file.
- Clone the repository to your local device.
- Once you open the source code directory, run the terminal and type the following in the terminal -> python manage.py migrate
- After migration is done type -> python manage.py runserver
- You will get the link to the web app local host, follow the link to run the web app.
However on some code-editors such as Visual-studio code, one may encounter an error as detailed below which is due to some of the visual studio code redistributables, as we encountered it, this occured on a Windows OS device. we strongly recommend the use of a virtual environment on developing and testing this project:
Could not find the DLL(s) 'msvcp140.dll or msvcp140_1.dll'. TensorFlow requires that these DLLs be installed in a directory that is named in your %PATH% environment variable. You may install these DLLs by downloading "Microsoft C++ Redistributable for Visual Studio 2015, 2017 and 2019" for your platform from this URL: https://support.microsoft.com/help/2977003/the-latest-supported-visual-c-downloads
Futhermore, for Linux OS users, one may get an error illegal instruction(core-dumped) when importing tensorflow as tf to solve the error one would need to downgrade the tensorflow cpu by running the following commands on the terminal:
pip uninstall tensorflow
pip install tensorflow==1.5
This will downgrade the cpu to device capability and prevent the app from crashing!!