I am training on pre-processed APTOS 2019 dataset for disease grading and the last two columns of my confusion matrix are constantly zero every time. Data is distributed as class0-1805 images, class1-370 images, class3-999 images, class4-193 images, class5-295 images. Here are results on 70 epochs:
Test acc=77%
Confusion matrix
[[258 3 1 0 0]
[ 6 38 16 0 0]
[ 13 13 130 0 0]
[ 0 7 22 0 0]
[ 4 9 30 0 0]]
Classwise accuracy on test set of model [98.47328244 63.33333333 83.33333333 0. 0. ]
Sensitivity (Recall) for each class: [0.98473282 0.63333333 0.83333333 0. 0. ]
Specificity for each class: [0.98513011 0.95416667 0.92592593 0.94727273 0.92181818]
using class weights didn’t help.
Please help!