Fig. 10From: Deep 3D mesh watermarking with self-adaptive robustnessBit accuracy rate (%) under different attacks and intensities on remeshed Asiadragon (top) and remeshed Hand (bottom) dataset. Model (red), Model\(\dag\) (cyan), Pre-Model (blue) and Pre-Model\(\dag\) (black) are showed. \(\dag\) denotes training without the attack layers, Model represents the model trained on the remeshed dataset, and Pre-Model represents the model trained on the original datasetBack to article page