Vinidra Diagnosis

The vision you deserve

This is a platform to diagnose TB using CNN.

• This project aims to create a disease diagnosis system by analysing medical images
• This perticular prototype aims to diagnose Tuberculosis using X-ray images
• We are using Tensorflow(Convolution neural networks) to predict Tuberculosis
• We used TBX11K(http://mmcheng.net/tb/) dataset to train this model
• We classify the result into healthy, sick(potential TB), and Active Tuberculosis
• We have deployed our model in Django and added some further functionalities
• We aim to extend it further by adding other diagnosis techniques and support for other diseases
• We aim to keep this project opensource to further extend this platform

Files overview

• tb_diagnosis/Diagnosis_model_code.ipynb: The actual model code
• tb_diagnosis/Model_check.ipynb: Checking the deployed model
• tb_diagnosis/install.py: Run this python file to install dependencies and download the Deep learning model in required path

Setup

Run the script tb_diagnosis/install.py to install dependencies and download the Deep learning model in required path

foo@bar:~$ cd ./tb_diagnosis
foo@bar:~$ python3 install.py

This is our model in action

https://www.youtube.com/watch?v=_l-XogdJYJ0&t=5s&ab_channel=AnuragTimilsina