Table 5 Performance comparison of proposed and state-of-the-art models in bone cancer classification.
References | Study | Methodology | Dataset/modality | Accuracy (%) | Precision | Recall | F1-score | ROC-AUC |
|---|---|---|---|---|---|---|---|---|
Vandana & Sathyavathi (2021) | Deep learning with image processing | Histopathology | 92.0 | 0.91 | 0.92 | 0.91 | 0.93 | |
Anisuzzaman et al. (2021) | CNNs (Inception V3, VGG19) | Histology | 96.0 | 0.95 | 0.96 | 0.95 | 0.96 | |
Punithavathi & Madhurasree (2023) | Extended CNN with wavelet-based segmentation | Histopathology | 97.0 | 0.96 | 0.97 | 0.96 | 0.97 | |
Alsubai et al. (2024) | GTOADL-ODHI with GF preprocessing, CapsNet, and SA-BiLSTM | Histopathological images | 97.5 | 0.97 | 0.97 | 0.97 | 0.98 | |
Ahmed et al. (2021) | Compact CNN model with oversampling | Histopathology | 96.8 | 0.96 | 0.96 | 0.96 | 0.97 | |
Alabdulkreem et al. (2023) | InceptionV3 and LSTM-based OSADL-BCDC | X-Ray | 95.0 | 0.94 | 0.95 | 0.94 | 0.95 | |
Proposed | Enhanced EfficientNet-B4 with Explainable AI | Transfer learning with Grad-CAM, SHAP, and Enhanced Bayesian Optimization | Histopathology | 97.9 (Binary)97.3 (Multi-Class) | 0.98 (Binary)0.97 (Multi-Class) | 0.98 (Binary)0.97 (Multi-Class) | 0.98 (Binary)0.97 (Multi-Class) | 0.99 (Binary)0.98 (Multi-Class) |
