Table 10 Comparative analysis of various deep learning models on 3D reconstruction under adversarial conditions using FGSM.
Model | HD | ED |
|---|---|---|
GAN-LSTM-3D10 | 2.585 ± 0.126 | 2.965 ± 0.056 |
YOLOv487 | 2.756 ± 0.220 | 2.856 ± 0.174 |
GAN-ResNet-3D83 | 2.826 ± 0.236 | 3.441 ± 0.120 |
GAN-GK-LSTM84 | 3.412 ± 0.142 | 3.268 ± 0.110 |
ViT88 | 2.048 ± 0.102 | 2.542 ± 0.142 |
GAN-LSTM-RL2 | 1.932 ± 0.154 | 2.206 ± 0.126 |
GAN-SP-LSTM89 | 1.862 ± 0.025 | 1.936 ± 0.068 |
Proposed | 0.933 ± 0.068 | 1.641 ± 0.053 |