Existing three-dimensional (3D) anomaly detection approaches typically rely on reconstruction, external memory banks, or fixed-radius clustering and often fail to generalize to noisy, irregularly sampled industrial scans or to capture the full diversity of real defects. We present Vote3D-AD, a single-pass framework that trains only on defect-free data and addresses these gaps with two principal contributions. First, we introduce Varied Defect Synthesis (VDS), a saliency-guided pseudo-anomaly generator that produces diverse, physically plausible defects (bulges, dents, holes, cracks, and surface roughness) together with sensor-level degradations to narrow the synthetic-to-real gap. Second, we develop a vote-and-cluster architecture in which local geometric representations predict learned, scale-aware votes that encode both spatial and boundary cues, and a differentiable soft-assignment clustering module aggregates these votes into coherent anomaly regions without relying on fixed-radius grouping or external memory structures. We evaluated our method on the synthetic Anomaly-ShapeNet benchmark and a new industrial dataset using three metrics: point-level Area Under the Receiver Operating Characteristic curve (AUROC), object-level Area Under the Precision-Recall curve (AUPR), and F1-Score. On average across both benchmarks, our method improves point-level AUROC by 6.7%, AUPR by 10.1% and F1 by 11.2%, and improves object-level AUROC by 5.3%, AUPR by 3.8% and F1 by 5.4% over the strongest baseline, while maintaining inference speeds above 9 frames per second (FPS).

In this study, we formulated a model that unifies the influence of media functions and information intervention on the transmission dynamics of the epidemic model. The media is used to inform people about the outbreak of a disease in a given population. The mass media function affects the rate of disease transmission by improving public awareness and promoting protective actions through the media. Information density reflects how much knowledge about the epidemic exists in the population, which determines individual responses and resource allocation over time. In this paper, we incorporate the mass media function and information density into the SIRS epidemic model. Our model provides a better perspective on the synergistic interplay of disease transmission, media-induced behavior changes, and the spread of information within the population. To explore the dynamics of the new model, the boundedness, non-negativity, and equilibria are analyzed. The local and global stability analyses of the new model are explored at feasible equilibrium points. The sensitivity analysis of the basic reproduction number and the epidemic equilibrium point is analyzed using a novel numerical method. Bifurcation analysis, such as saddle node, transcritical, pitchfork, forward, and backward bifurcations, has been carried out to highlight the most prominent qualitative changes in the model dynamics. Subsequently, we extended an approach known as disease-informed neural networks (DINNs), which can be applied to efficiently project the spread of diseases and recovery. The application of physics-informed neural networks (PINNs) is formulated on the proposed model. We expound on how neural networks have the potential to learn the degree of spread of the epidemic in a society: prediction of evolution, exploring some sensitive parameters. To demonstrate the strength and usefulness of DINNs, some computational experiments will be performed, including Gaussian noise to synthetic data, and the results proved that DINNs is a robust approach to efficiently estimate the dynamics of the disease spread and project future behavior based on synthetically generated data.