Table 3.
Comparison of the pre-trained models using transfer learning technique with existing state-of-the-art artworks.
| References | Methodology | Dataset | Accuracy (%) |
|---|---|---|---|
| Zhang et al. (2013) | DWT + PCA + PSO optimized Kernel SVM | 90 images | 97.78 |
| Yang et al. (2016) | DWT + BBO optimized SVM | 90 images | 97.78 |
| Lu et al. (2017) | DWT + BAT optimized machine learning | 132 images | 98.33 |
| Nayak et al. (2018) | Ripplet-II Features + PCA + ModifiedPSO based ELM | 160 images | 99.26 |
| Zhang et al. (2018) | Pseudo Zernike moment + Kernel SVM | 160 images | 99.45 |
| Khan et al. (2020) | KM clustering + Deep learning with data augmentation | 220 images | 94.06 |
| Proposed work | VGG-19 with Transfer Learning | 305 images | 99.48 |
| VGG-16 with Transfer Learning | 99 | ||
| ResNet50 with Transfer Learning | 97.92 | ||
| Inception V3 with Transfer Learning | 81.25 |
BBO, Biogeography based optimization; DWT, Discrete Wavelet Transform; ELM, Extreme Learning Machine; PCA, Principal Component Analysis; PSO, Particle swarm optimization; SVM, Support Vector Machine.