Table 1 Limitations of related work.

From: Intelligent skin disease prediction system using transfer learning and explainable artificial intelligence

Author

Image based Dataset

Technique

Outcomes

Limitations

Ali et al.47

Chickenpox, measles, and monkeypox

VGG16, ResNet50, InceptionV3, and Ensemble

Achieved accuracies of 81.48 (± 6.87%), 82.96 (± 4.57%), 74.07 (± 3.78%), and 79.26(± 1.05%).

(i) Low accuracy

(ii) No use of explainable artificial intelligence

Burak Gülmez48

Chickenpox, measles, monkeypox, and normal

Ensemble approach

Achieved an accuracy of 84.2%.

(i) Low accuracy

(ii) Small dataset

(iii) No use of data augmentation

(iv) No use of explainable artificial intelligence

Irmak et al.49

Chickenpox, measles, monkeypox, and normal

MobileNetV2, VGG16, and VGG19

Achieved accuracies of 91.38%, 83.62%, and 78.45%.

(i) Low accuracy

(ii) Small dataset

(iii) No use of data augmentation

(iv) No use of explainable artificial intelligence

Singh and Songare50

Monkeypox and normal

VGG-16, ResNet50, InceptionV3, and GoogLeNet

Achieved accuracies of 83.85%, 85.38%, 86.37%, and 88.27%.

(i) Low accuracy

(ii) Less classes

(iii) No use of explainable artificial intelligence

Sharma et al.51

Monkeypox, measles, and chickenpox

ResNet18-based model

Achieved an accuracy of 84.59%.

(i) Low accuracy

(ii) Small dataset

(iii) No use of explainable artificial intelligence

Sethy et al.52

Chickenpox, measles, monkeypox, and normal

Darknet 19 and Improved Darknet 19

Achieved accuracies of 81.4% and 85.49%.

(i) Low accuracy

(ii) Small dataset

(iii) No use of data augmentation

(iv) No use of explainable artificial intelligence

Uysal53

Chickenpox, measles, monkeypox, and normal

CNN-LSTM hybrid model

Achieved an accuracy of 87%.

(i) Low accuracy

(ii) No use of explainable artificial intelligence

Ariansyah et al.54

Monkeypox, measles, and normal

CNN and VGG16

Achieved accuracies of 64.52% and 83.33%.

(i) Low accuracy

(ii) Small dataset

(iii) No use of data augmentation

(iv) No use of explainable artificial intelligence

Kundu et al.55

Chickenpox, measles, and monkeypox

SVM, KNN, RestNet50, and ViT

Achieved accuracies of 65%, 84%, 91%, and 93%.

(i) Low accuracy

(ii) No use of explainable artificial intelligence

Aqsa Akram et al.56

Chickenpox, measles, monkeypox, and normal

SkinMarkNet

Achieved an accuracy of 90.615%.

(i) Low accuracy

(ii) No use of explainable artificial intelligence

Back to article page