Table 6.
Comparison of various research studies along with our proposed model.
| References | Images | Model | Sensitivity | Specificity | AUC score | Accuracy |
|---|---|---|---|---|---|---|
| Liu et al. [2] | 4701 | Alex net and Google net | — | — | — | 85.68 |
| Hooda et al. [27] | 1133 | Alex net and Google net and ResNet ensemble | 88.42 | 88 | 93 | 88.24 |
| Liu et al. [26] | 11200 | SSD without pretraining and VGG net 1–16 as back bone | 88.4 | 89.5 | 93.8 | 88.2 |
| Hwang et al. [24] | 10848 | Modified pretrained Alex net | — | — | 96.4 | 90 |
| Ghorakavi et al. [28] | 800 | Reset 18 with data augmentation | — | — | — | 65.771 |
| Heo et al. [29] | 10000 | D-CNN | 81.5 | 96.2 | 92 | — |
| Lakhani and Sundaram [30] | 1007 | Ensemble of Alex net and Google Net | 97.3 | 94.7 | 99 | 96 |
| Nguyen et al. [31] | 800 | DenseNet-121 | — | — | — | 80 |
| Sivaramakrishnan et al. [32] | 112,782 | Inception v3 | 72 | 82 | 70.54 | 99 |
| Hijazi et al. [33] | 800 | VGG 16 and inception V3 | 90.91 | 88.64 | — | 89.77 |
| Pasa et al. [34] | 1104 | CNN | — | — | 92.5 | 86.2 |
| Jaeger et al. [35] | 135 | ANN,CNN, VGG16 | — | — | 66 | — |
| Vajda et al. [36] | 814 | Neural network | — | — | 99 | 97.03 |
| Lopes et al. [37] | 1120 | Gooogle net Res Net VGG net SVM | — | — | 92.6 | 84.7 |
| Proposed Model | 3500 | NF net model | 91.81 | 98.42 | 99.38 | 96.91 |