Due to the complex behavior of clothing, modeling fine-scale garment deformation on arbitrary meshes within a unified network presents a considerable challenge. Current methods often fail to learn generalized patterns that adhere to physical laws in a controllable manner, compromising efficiency and realism in practical applications. To overcome these limitations, we introduce a novel hybrid learning approach that accurately simulates garment dynamics and intricate details. The core of our method is a progressive learning scheme that integrates the accuracy of supervised learning with the physical awareness derived from unsupervised learning. Additionally, after analyzing the nature of the garment self-collision problem, we introduce vertex repulsion-based constraints that effectively prevent conspicuous intersections within the mesh while preserving fine details. Finally, to ensure effective propagation of node information across the mesh, we propose a multi-scale graph processing technique for the garment deformation task, featuring structure-preserving pooling and unpooling strategies that significantly enhance result quality. The experimental results demonstrate that our method outperforms the state-of-the-art under multiple metric evaluations. © 2025 Elsevier B.V.

Despite significant advancements in neural simulation techniques for clothing animation, these methods struggle to capture the dynamic details of garments during movement. This limitation restricts their applicability in scenarios where high-quality garment deformation is essential. To address this challenge, we introduce a novel graph learning-based approach to enhance deformation realism through designed mechanisms for mesh information propagation and external optimization strategies during model training. First, we address the issue of over-smoothing common in conventional graph processing techniques by introducing a flexible message-passing method. This approach effectively manages node interactions within the mesh, thereby improving the expressiveness of the model. Furthermore, acknowledging that uniform model supervision typically neglects high-frequency details during optimization, we analyze the spectral properties of clothing meshes. Based on this analysis, we introduce a frequency-division constraint aligned with the characteristics of different frequency bands, which aids in precisely controlling the generation of details. Our model further integrates self-collision and other physics-aware losses, enabling the learning of generalized and fine-grained dynamic deformations. Extensive evaluations and comparisons demonstrate the effectiveness of our approach, showing notable improvements over existing state-of-the-art solutions.

We present a novel solution for mesh-based deformation simulation from a spectral perspective. Unlike existing approaches that demand separate training for each garment or body type and often struggle to produce rich folds and lifelike dynamics, our method achieves the quality of physics-based simulations while maintaining superior efficiency within a unified model. The key to achieve this lies in the development of a spectrum-enhanced deformation network, a result of in-depth theoretical analysis bridging neural networks and garment deformations. This enhancement compels the network to focus on learning spectral information predominantly within the frequency band associated with intricate deformations. Furthermore, building upon standard blend skinning techniques, we introduce target-aware temporal skinning weights. The weights describe how the underlying human skeleton dynamically affects the mesh vertices according to the garment and body shape, as well as the motion state. We validate our method on various garments, bodies, and motions through extensive ablation studies. Finally, we conduct comparisons to confirm its superiority in generalization, deformation quality, and performance over several state-of-the-art methods.