Mobile Information Systems / 2022 / Article / Tab 8 / Research Article
Hyper-Tuned CNN Using EVO Technique for Efficient Biomedical Image Classification Table 8 Performance of evaluation of traditional ANN-based strategies, DNN based strategies, and the proposed hyperparameter tuned CNN based on EVO for the Glioma dataset.
Models Class Accuracy Precision Recall Specificity F1-score MCR FDR TNR MLP Glioma tumour 0.8372 0.6731 0.7000 0.8840 0.6863 0.1628 0.3269 0.3 Meningioma tumour 0.8144 0.6612 0.7207 0.8520 0.6897 0.1856 0.3388 0.2793 No tumour 0.8663 0.7609 0.7000 0.9239 0.7292 0.1337 0.2391 0.3 Pituitary tumour 0.9049 0.7463 0.7143 0.9467 0.7299 0.0951 0.2537 0.2857 ELM Glioma tumour 0.8878 0.7407 0.8333 0.9054 0.7843 0.1122 0.2593 0.1667 Meningioma tumour 0.8817 0.7627 0.8333 0.9004 0.7965 0.1183 0.2373 0.1667 No tumour 0.9302 0.8989 0.8163 0.9689 0.8556 0.0698 0.1011 0.1837 Pituitary tumour 0.9308 0.8000 0.8116 0.9564 0.8058 0.0692 0.2 0.1884 DNN Glioma tumour 0.9412 0.8700 0.8969 0.9558 0.8832 0.0588 0.13 0.1031 Meningioma tumour 0.9487 0.9091 0.9091 0.9643 0.9091 0.0513 0.0909 0.0909 No tumour 0.9523 0.9000 0.9000 0.9687 0.9000 0.0477 0.1 0.1 Pituitary tumour 0.9512 0.8696 0.8571 0.9718 0.8633 0.0488 0.1304 0.1429 CNN Glioma tumour 0.9820 0.9574 0.9677 0.9864 0.9626 0.018 0.0426 0.0323 Meningioma tumour 0.9747 0.9402 0.9735 0.9752 0.9565 0.0253 0.0598 0.0265 No tumour 0.9844 0.9890 0.9474 0.9966 0.9677 0.0156 0.011 0.0526 Pituitary tumour 0.9746 0.9333 0.9333 0.9843 0.9333 0.0254 0.0667 0.0667 Hyperparameter tuned CNN Glioma tumour 0.9847 0.9691 0.9691 0.9898 0.9691 0.0153 0.0309 0.0309 Meningioma tumour 0.9847 0.9735 0.9735 0.9892 0.9735 0.0153 0.0265 0.0265 No tumour 0.9923 0.9901 0.9804 0.9965 0.9852 0.0077 0.0099 0.0196 Pituitary tumour 0.9873 0.9722 0.9589 0.9938 0.9655 0.0127 0.0278 0.0411