Skip to content

This notebook has the model created to predict COVID-19 only based on X-ray chest scans.

Notifications You must be signed in to change notification settings

sebassaras02/COVID-19-CNN-ResNet152-V2

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

Convolutional Neural Network based on ResNet152-V2 to predict COVID-19 disease

This repository contains the Convolutional Neural Network (CNN) model created to predict the COVID-19 disease based only in the x-ray chest scans.

Table of Contents

  • covid-19-transfer-learning-resnet152v2.ipynb is the jupyter notebook with the data treatment, model creation and model validation.

Data


Model results


The best model was the one who used transfer learning with a ResNet152-V2 architecture. The model was able to get 0.65 of loss and 0.92 of accuracy.

FlowChart of the Model Construction


  graph TD
    A[Load data] --> B(Split into train, validation, and test)
    B --> C{Create a big CNN }
    C --> D[Validate with test data]
    D -->|Yes| E[Export the model]
    D -->|No| F[Deal with overfitting]
    F --> G[Create a simpler model]
    G --> H[Validate the simpler model]
    H -->|Yes| I[Export the model]
    H -->|No| J[Deal with underfitting]
    J --> K[Use transfer learning]
    K --> L[Add new layers]
    L-->|Yes| M[Export the model]
    L-->|No| N[Repeat all again]
    

Loading

Contacts:


About

This notebook has the model created to predict COVID-19 only based on X-ray chest scans.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published