Personal Projects

EMCNet: Automated Covid-19 Diagnosis

Implementing a Paper titled "Automated COVID-19 diagnosis from X-ray images using CNN and ensemble of ML classifiers" since this paper combined some interesting concepts. That was a great opportunity to learn and test myself in learning theoretical concepts and practical skills.
Also I designed a GUI for more convenient use such that users can add their X-ray pictures and the model will give them the class with confidence percentage.
All the codes and results are available in my github. For summary some of them have been shown in next section:

Train & Validation Loss and Accuracy for paper architecture without any changes

Train & Validation Loss and Accuracy considering Mean & Std of pictures and applying it

SVM

Confusion Matrices

for SVM & Adaboost Classifiers

Both of them classified all data similarly

Adaboost

Random Forest

Confusion Matrices

for Random Forest & Decision Tree Classifiers

Random Forest performed better than Decision Tree

Decision Tree

Ensemble Classifier

Confusion Matrices

for Ensemble Classifiers
&
Receiver Operating Characteristic plot
which are similar to paper

ROC plot matches the Paper's ROC plot

Shallow convolutional neural network for image classification

Implementing a Paper titled "Shallow convolutional neural network for image classification" which was a turning point in computer vision introduced a simple high-efficiency CNN architecture to improve performance and computing reduction.

Page updated

Google Sites

Report abuse