Table 1.
Publication | Model | Accuracy | Dataset | Limitations |
---|---|---|---|---|
Swain et al. (4) | SVM | 98.13% | Private | • Dataset is small |
Nanglia et al. (5) | KNN+ SVM + DT | 78% | Private | • Less accuracy • Required hand crafted features |
Krizhevsky et al. (7) | SVM | 88% | Private | • Required hand crafted features • Dataset is not available publically |
Dhabyani, et al. (10) | AlexNett+ VGG 16+ Inception+ Res net+ Nasnet | 78%, 88%, 85%, 93%, 94%, 80%, 82%, 80%, 90%, 92% | Public | • Less accuracy |
Spanhol et al. (11) | SVM | 80% | Private | • Less accuracy • Required hand-crafted features |
Yap et al. (12) | AlexNet | 91%, 89% | Private | • Use of imbalanced dataset • Dataset is small |
Saber et al. (14) | Inception v3 + Resnet 50+ VGG 16+ VGG 19 | 96%, 94%, 96%, 95% | Public | • Dataset is small • Lack of diverse dataset |