Table 3 Performance comparison of the proposed model and existing models. Bold texts represent the proposed model.
From: DANet a lightweight dilated attention network for malaria parasite detection
Model | ACC (%) | PR (%) | SE (%) | SP (%) | F1 (%) | Parameter Count (M) |
|---|---|---|---|---|---|---|
EfficientNetB026 | 94.45 | 92.85 | 95.91 | 93.07 | 94.36 | 5.30 |
ResNet15226 | 94.29 | 91.15 | 97.25 | 91.67 | 94.10 | 115.6 |
NASNetMobile26 | 93.21 | 92.34 | 93.98 | 92.48 | 93.16 | 5.33 |
InceptionV326 | 93.02 | 90.46 | 95.33 | 90.92 | 92.83 | 25.00 |
InceptionResNetV226 | 94.03 | 91.87 | 96.02 | 92.21 | 93.90 | 55.00 |
Yang F, et al.27 | 93.46 | 82.73 | 94.25 | 98.39 | 80.81 | – |
VGG1628 | 95.62 | 97.28 | 92.17 | 65.02 | 94.66 | 1383 |
VGG1928 | 96.31 | 95.75 | 95.02 | 63.49 | 95.39 | 1436 |
Detailed CNN29 | 97.24 | 96.10 | – | 98.47 | 97.24 | 16.79 |
Qayyum A. et al.30 | 96.05 | 95.80 | 96.33 | – | 96.06 | – |
ResNet5024 | 88.47 | – | 89.61 | 87.31 | 88.57 | 23.6 |
DenseNet12131 | 90.94 | – | 92.51 | 89.33 | 91.03 | 7.04 |
DPN9232 | 87.88 | – | 86.81 | 88.98 | 87.85 | 37.7 |
DCNN(Falcon)-TL33 | 93.00 | 93.00 |  |  | 93.00 | – |
VGG16-SVM22 | 93.01 | 84.47 | 94.10 | 94.90 | 87.05 | 1383 |
Mosquito-Net24 | 96.60 | 93.00 | 97.60 | 95.80 | 96.70 | 7.47 |
Alex-Net24 | 92.70 | 93.20 | 93.90 | 93.10 | 93.90 | 61.1 |
Xception-Net24 | 88.90 | 93.00 | 94.10 | 84.10 | 88.90 | 22.8 |
DLRFNet34 | 94.32 | 94.68 |  |  | 93.78 | – |
CNN-LSTM-BiLSTM35 | 96.20 | 97.00 | – | – | 97.00 | – |
Hybrid CapsNet36 | 99.07 | 97.60 | – | 97.30 | 97.40 | – |
97.36 | – | 97.46 | 95.00 | 97.39 | – | |
MosquitoNet38 | 96.97 | – | – | – | – | – |
CNN Model39 | 95.00 | 94.50 | – | – | 94.00 | – |
Transformer-CNN | 94.63 | 91.48 | 98.32 | 94.77 | 27.66 | Â |
DANet-H+LogSoftMax | 97.65 | 97.45 | 97.93 | 97.34 | 97.69 | 2.3 |
DANet-S+Sigmoid | 97.75 | 98.89 | 96.65 | 96.60 | 97.73 | 2.3 |
DANet-S+LogSoftMax | 97.95 | 97.76 | 98.07 | 97.87 | 97.86 | 2.3 |
