Inspiration
We were inspired to create this project after learning that a startling number of diseases are diagnosed incorrectly. This has only increased during the COVID pandemic, especially with pneumonia. Our classifier aims to reduce this issue by giving medical professionals a second source of input on whether the condition is pneumonia or not.
What it does
The model predicts whether the patient has pneumonia or not based on an x-ray of the chest.
How we built it
We built the deep learning using TensorFlow and we used a few others like NumPy for preprocessing.
Challenges we ran into
We struggled with the preprocessing of the images. This took up the majority of the time we spent working on the project.
Accomplishments that we're proud of
We're proud of getting the accuracy of the model above 90%
What we learned
We learned about TensorFlow image generators which we needed to use to batch the data and turn the NumPy arrays of the images into a format the model could use to train.
What's next for Pneumonia Classifier
We will make a streamlit web app so users can directly upload an x-ray and get a prediction
Log in or sign up for Devpost to join the conversation.