ParuParuKu
ParuParuKu is a Flask-based web application for detecting pneumonia from X-ray images that helps patients detect pneumonia early.
ParuParuKu uses the best 3 out of 5 models that have been compared to provide users with highest accuracy prediction as possible.
Problem Statement:
Pneumonia is an inflammation of the tissue in lungs caused by infection. Pneumonia can cause mild to severe symptoms, such as coughing, fever, nausea, wheezing, shortness of breath, and chest pain. Pneumonia is not only found in adults but also in children.
There are times when doctors aren't available to make a diagnosis on recently performed X-ray scans. So early diagnosis is needed to take further action.
Challenges:
Help patients to have early pneumonia diagnosis using Artificial Intelligence.
Flowchart:
Site Map:
Since this application provides multiple languages, the sitemap provided is as below.
Technologies:
Languages
- HTML (Layout)
- Python
Frameworks / Libraries
Machine Learning Models
- Tensorflow Keras
- Scikit-learn
- Numpy
- Seaborn (visualization purpose)
- Matplotlib (visualization purpose)
- Gdown (download sorted large datasets from google drives)
Flask Web Application
- Flask
- Flask Babel
- Tensorflow-CPU
- Pillow
- Gunicorn
- Werkzeug
- Scipy
- OpenCV
Defined in Requirements.txt for docker container.
Other Framework
- Bootstrap
Tools
- Google Colaboratory
- Visual Studio Code
- Figma
Screenshots:
Others:
Sidenotes
ParuParuKu was submitted and awarded as 2nd winner of The 1st International Engineering Festival hosted by Universitas Internasional Batam. Detail submission can be found on this link.
Presentation
Requirements.txt
Credits/Shoutouts:
This project were developed in the team of 4, consists of
- Henry
- Marvin Christian
- Nicky Nicholas
- And me
Also credits towards