Table 8 Comparison with existing methods
From: Tongue image quality assessment based on a deep convolutional neural network
Model | Precision | Recall | F1-score | Accuracy |
|---|---|---|---|---|
SVM | 83.39% (7.00%) | 83.25% (4.32%) | 83.06% (3.23%) | 82.95% (3.20%) |
Decision Tree | 79.52% (4.41%) | 76.39% (6.74%) | 77.66% (3.86%) | 78.09% (2.86%) |
Random Forest | 84.10% (6.76%) | 81.67% (4.51%) | 82.64% (3.70%) | 82.84% (3.35%) |
Naïve Bayes | 88.86% (6.15%) | 61.24% (7.15%) | 72.26% (5.88%) | 76.57% (4.39%) |
Adaboost | 83.50% (4.60%) | 79.23% (6.70%) | 81.00% (2.74%) | 81.42% (2.26%) |
GBDT | 84.46% (4.82%) | 81.72% (5.61%) | 82.86% (3.36%) | 83.15% (2.79%) |
VGG-16 | 91.87% (2.14%) | 91.68% (2.16%) | 91.77% (2.15%) | 91.68% (2.16%) |
ResNet-152 | 96.36% (2.81%) | 96.32% (2.81%) | 96.34% (2.81%) | 96.32% (2.81%) |
DenseNet-169 | 96.30% (2.19%) | 96.22% (2.23%) | 96.26% (2.21%) | 96.22% (2.23%) |