Table 1.
Recent studies employing computer vision algorithms for species classification across various taxonomic groups.
| Species | Samples | Architecture | Accuracy | Study |
|---|---|---|---|---|
| Reptiles | 386,006 | Vision Transformer (ViT) | 0.962 | Bolon et al. (2022) |
| Reptiles | 82,601 | EfficientNet | 0.870 | Durso et al. (2021) |
| Lizards & Amphibians | 6,045 | MobileNetV2 | 0.820 | Gill et al. (2024) |
| Lizards & Amphibians | 2,700 | VGG16 | 0.870 | Binta Islam et al. (2023) |
| Fishes | 1,080 | Image Processing + SVM | 0.942 | Sharmin et al. (2019) |
| Fishes | 3,068 | U-NET + CNN | 0.979 | Robillard et al. (2023) |
| Lizards & Amphibians | 828 | CNN | 0.600 | Islam and Valles (2020) |
| Mammals | 326 | Mask R-CNN + ResNet101 | 0.980 | Gray et al. (2019) |
A total of seven comparable studies where pre-trained state-of-the-art deep learning models were employed on taxonomic datasets of different groups were analyzed in order to provide more robustness to our research. The sample size and accuracy score are reported below for each study.