Skip to content

This project is about the creation of an AI: The architecture is a CNN with FFC, also implements an U-NET architecture. Its main function is, given a chest-radiography image, be able to diagnostique pulmonar damage (pneumonia) by the recognition of inflamatory pulmonar tissue.

Notifications You must be signed in to change notification settings

AlanAmaro13/IA_U-NET_FOR_PNEUMONIA_DETECTION

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hi! I'm Alan Amaro and this is a final project that I make for a college class: TSFC-I: Inteligencia Artificial en la Fisica.

You can see three folders, the notebook's folders contain all the main code source for image analysis and processing, and the central notebook contains all the architecure for the CNN and the U-Net architectures. The Best_Model folder contains only the model in .h5 and .keras extensions, this model has done with only a simple CNN and a FFC.

The U-Net model is the 'best model' that I was able to create, this implements and U-Net architecture in .h5 and .keras extensions. The image is process by the U-Net, and later is given to the first architecture (CNN-FFC). Both models have an acurracy above 90% (U-Net has 96%)

All the data was taken for kaggle.com, in the first notebook I share the link to all data.

Have fun watching or implementing this code to your project!

About

This project is about the creation of an AI: The architecture is a CNN with FFC, also implements an U-NET architecture. Its main function is, given a chest-radiography image, be able to diagnostique pulmonar damage (pneumonia) by the recognition of inflamatory pulmonar tissue.

Topics

Resources

Stars

Watchers

Forks