I have created this project for clssify the ct scan images of covid19, pneumonia ,and normal. I used vgg19 + 2 more layers of cnn to classify this images train on different different epoch
You can find the dataset that i used link here
The architecture of model are :
the model result on train and validation as following :
![]() |
![]() |
![]() |
![]() |
You can find the wandb link here
heres i got 99% AUC and RECall for the training data but 62% recall and 75% AUC for the validation data .this is the first model that i used on this dataset i am try to build some more architecture and tuning the hyperperameter so i can get a good recall and AUC for the validation dataset.