Table 5 Comparison of proposed method and related literature
Method | Subject | Number of images | Classifier | Accuracy (%) |
|---|---|---|---|---|
Rajendra et al. [37] | 20 subjects age: 29 to 74 | 2600 (1300 benign/13000 malignant) 130 images from each subject | KNN | 100% |
Rajendra et al. [38] | 469 subjects age: 23 to 90 | 469 (238 benign/231 malignant) | fuzzy forest | 80.6% |
Wu et al. [39] | 988 subjects age: 7 to 84 | 988 (the number of images in each class were not provided) | CNN (DenseNet) | 87.5% |
Our proposed method | 587 subjects age: 11 to 85 | 1896 (948 benign/948 malignant) | Ensemble CNN | 92.15% |