Table 6 Classification results obtained when the models are run 1 vs. 1 (5 slices manually).
1 vs. 1 | Models | Accuracy | Precision | Recall | F1 Score | AUC | MCC |
|---|---|---|---|---|---|---|---|
CN/AD | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9756 0.9012 0.9858 0.9128 0.9881 0.9859 | 0.9473 0.7627 0.9673 0.7922 0.9718 0.9645 | 0.9423 0.8194 0.9612 0.7582 0.9688 0.9640 | 0.9446 0.7927 0.9670 0.7757 0.9720 0.9646 | 0.9838 0.9719 0.9882 0.9591 0.9982 0.9861 | 0.9491 0.8375 0.9702 0.8236 0.9734 0.9668 |
CN/MCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9649 0.9283 0.9891 0.9172 0.9872 0.9788 | 0.9448 0.8289 0.9718 0.7983 0.9696 0.9469 | 0.9429 0.8194 0.9799 0.7681 0.9664 0.9469 | 0.9440 0.8267 0.9731 0.7824 0.9685 0.9469 | 0.9819 0.9519 0.9995 0.9675 0.9982 0.9711 | 0.9474 0.8299 0.9754 0.7903 0.9698 0.9469 |
CN/LMCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9783 0.9192 0.9896 0.9189 0.9881 0.9898 | 0.9492 0.8122 0.9746 0.7997 0.9715 0.9797 | 0.9413 0.7895 0.9735 0.7817 0.9691 0.9794 | 0.9445 0.8019 0.9742 0.7906 0.9745 0.9795 | 0.9962 0.9691 0.9987 0.9781 0.9818 0.9822 | 0.9507 0.8008 0.9741 0.7923 0.9717 0.9795 |
CN/EMCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9769 0.8983 0.9883 0.9247 0.9850 0.9892 | 0.9428 0.7712 0.9718 0.8283 0.9650 0.9728 | 0.9438 0.7912 0.9683 0.8012 0.9600 0.9382 | 0.9427 0.7812 0.9742 0.8140 0.9628 0.9548 | 0.9849 0.9591 0.9819 0.9699 0.9875 0.9884 | 0.9431 0.7806 0.9714 0.8144 0.9626 0.9529 |
AD/MCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9819 0.9383 0.9881 0.9239 0.9900 0.9890 | 0.9592 0.8619 0.9891 0.8634 0.9864 0.9508 | 0.9521 0.8371 0.9721 0.7898 0.9737 0.9491 | 0.9540 0.8492 0.9795 0.8276 0.9798 0.9502 | 0.9920 0.9781 0.9961 0.9695 0.9986 0.9818 | 0.9567 0.8494 0.9820 0.8269 0.9800 0.9500 |
AD/LMCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9449 0.9693 0.9942 0.9757 0.9945 0.9697 | 0.8544 0.9382 0.9809 0.9369 0.9868 0.9268 | 0.8529 0.9193 0.9891 0.9371 0.9856 0.9215 | 0.8515 0.9283 0.9799 0.9360 0.9850 0.9226 | 0.9733 0.9955 0.9973 0.9947 0.9988 0.9503 | 0.8529 0.9274 0.9850 0.9367 0.9861 0.9236 |
AD/EMCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9895 0.9482 0.9822 0.9683 0.9919 0.9475 | 0.9742 0.8910 0.9810 0.9099 0.9813 0.8741 | 0.9739 0.8618 0.9701 0.9074 0.9780 0.8614 | 0.9740 0.8783 0.9748 0.9072 0.9791 0.8687 | 0.9917 0.9891 0.9946 0.9889 0.9988 0.9261 | 0.9739 0.8763 0.9753 0.9082 0.9791 0.8671 |
MCI/LMCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9518 0.8973 0.9881 0.9181 0.9891 0.9887 | 0.9129 0.7672 0.9784 0.7896 0.9737 0.9621 | 0.9192 0.7221 0.9794 0.7724 0.9717 0.9612 | 0.9172 0.7435 0.9782 0.7797 0.9701 0.9626 | 0.9839 0.9719 0.9889 0.9691 0.9985 0.9808 | 0.9164 0.7442 0.9786 0.7806 0.9718 0.9619 |
MCI/EMCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9749 0.8771 0.9795 0.9164 0.9888 0.9862 | 0.9317 0.7219 0.9484 0.8091 0.9739 0.8496 | 0.9311 0.6329 0.9463 0.7793 0.9700 0.8278 | 0.9290 0.6736 0.9461 0.7934 0.9702 0.8393 | 0.9820 0.9219 0.9871 0.9655 0.9984 0.9540 | 0.9300 0.6610 0.9470 0.7933 0.9713 0.8360 |
LMCI/EMCI | EfficientNetB2 InceptionV3 RegNetx006 Basic ViT EfficientNetB2 + FPN (Our Model) ViT + EfficientNetB2 + FPN (Our Model) | 0.9839 0.9481 0.9871 0.9695 0.9846 0.9877 | 0.9744 0.8692 0.9883 0.9294 0.9771 0.9861 | 0.9753 0.8516 0.9845 0.9174 0.9756 0.9820 | 0.9741 0.8594 0.9871 0.9237 0.9771 0.9840 | 0.9912 0.9771 0.9713 0.9723 0.9893 0.9932 | 0.9746 0.8597 0.9866 0.9235 0.9763 0.9843 |
