A5102778414- 3D Shape Modeling and Analysis
- Computer Graphics and Visualization Techniques
- Advanced Numerical Analysis Techniques
- Computational Geometry and Mesh Generation
- Model Reduction and Neural Networks
- Infrared Target Detection Methodologies
- COVID-19 diagnosis using AI
- CCD and CMOS Imaging Sensors
- Elasticity and Material Modeling
- Civil and Geotechnical Engineering Research
- Radiomics and Machine Learning in Medical Imaging
- Human Pose and Action Recognition
- Advanced Vision and Imaging
- Dynamics and Control of Mechanical Systems
- Image Processing Techniques and Applications
- Fluid Dynamics Simulations and Interactions
- 3D Modeling in Geospatial Applications
- Lung Cancer Diagnosis and Treatment
- Human Motion and Animation
- Lattice Boltzmann Simulation Studies
- Grouting, Rheology, and Soil Mechanics
University of Utah
2023-2024
Clemson University
2021
University of New Mexico
2020
Northeastern University
2019
Xiamen University
2015-2017
Southwest Jiaotong University
2011
In this paper, we present a GPU algorithm for finite element hyperelastic simulation. We show that the interior-point method, known to be effective robust collision resolution, can coupled with non-Newton procedures and massively sped up on GPU. Newton's method has been widely chosen family, which fully solves linear system at each step. After that, active set associated collision/contact constraints is updated. Mimicking routine using optimization (like gradient descent or ADMM)...
3D Gaussian Splatting (3DGS) has emerged as a mainstream solution for novel view synthesis and reconstruction. By explicitly encoding scene using collection of kernels, 3DGS achieves high-quality rendering with superior efficiency. As learning-based approach, training been dealt the standard stochastic gradient descent (SGD) method, which offers at most linear convergence. Consequently, often requires tens minutes, even GPU acceleration. This paper introduces (near) second-order convergent...
We propose a framework for the interactive simulation of nonlinear deformable objects. The primary feature our system is seamless integration and collision culling, which are often independently handled in existing animation systems. bridge connecting them medial axis transform (MAT), high-fidelity volumetric approximation complex 3D shapes. From physics perspective, MAT leads to an expressive compact reduced model. employ semireduced projective dynamics formulation, well captures...
In recent years, the research into automatic aided detection systems for pulmonary nodules has been extremely active. Most of existing studies are based on 2D convolution neural networks, which cannot make full use computed tomography’s (CT) 3D spatial information. To address this problem, a computer-aided (CAD) system lung residual network (3D-ResNet) inspired by cognitive science is proposed in paper. system, we feed slice information extracted from three different axis planes U-NET set,...
This paper provides a new avenue for exploiting deep neural networks to improve physics-based simulation. Specifically, we integrate the classic Lagrangian mechanics with autoencoder accelerate elastic simulation of deformable solids. Due inertia effect, dynamic equilibrium cannot be established without evaluating second-order derivatives network. is beyond capability off-the-shelf automatic differentiation packages and algorithms, which mainly focus on gradient evaluation. Solving nonlinear...
We propose an efficient cloth simulation method that combines the merits of two drastically different numerical procedures, namely subspace integration and parallelizable iterative relaxation. show those methods can be organically coupled within framework projective dynamics (PD), where both low- high-frequency motions are effectively efficiently computed. Our works seamlessly with state-of-the-art contact handling algorithm, incremental potential (IPC), to offer non-penetration guarantee...
Abstract Researchers have proposed various techniques for cloth simulation in the last decades. The crucial problem interactive animation is how to speed up with a stable system. In this paper, we describe realistic and scheme based on mass‐spring This modifies semi‐implicit integration used system an efficient damping method. Semi‐implicit methods been widely dynamic because of acceptable stability. However, internal forces are generated respect rotational rigid motions, which only...
Mil2 pushes the performance of high-resolution cloth simulation, making simulation interactive (in milliseconds) for models with one million degrees freedom (DOFs) while keeping every triangle untangled. The guarantee being penetration-free is inspired by interior-point method, which converts inequality constraints to barrier potentials. Nevertheless, we propose a major overhaul this modality defining novel and simple formulation does not depend on distance between mesh primitives. Such...
This paper pushes the performance of cloth simulation, making simulation interactive even for high-resolution garment models while keeping every triangle untangled. The penetration-free guarantee is inspired by interior point method, which converts inequality constraints to barrier potentials. We propose a major overhaul this modality within projective dynamics framework leveraging an adaptive weighting mechanism formulation. approach does not depend on distance between mesh primitives, but...
This paper presents volumetric homogenization, a spatially varying homogenization scheme for knitwear simulation. We are motivated by the observation that macro-scale fabric dynamics is strongly correlated with its underlying knitting patterns. Therefore, towards single material less effective when complex and non-repetitive. Our method tackles this challenge homogenizing yarn-level locally at elements. Assigning virtual volume of structure enables us to model bending twisting effects via...
Position-based Dynamics (PBD) and its extension, eXtended (XPBD), have been predominantly applied to compliant constrained elastodynamics, with their potential in finite strain (visco-) elastoplasticity remaining underexplored. XPBD is often perceived stand contrast other meshless methods, such as the Material Point Method (MPM). MPM based on discretizing weak form of governing partial differential equations within a continuum domain, coupled hybrid Lagrangian-Eulerian method for tracking...
With the development and utilization of underground space construction large quantities shield tunnels, phenomenon that tunnel go through existing buildings at close quarters is in rise. The stiffness mismatch between tunnels has a significant impact on deformation stress connection part. In order to explore affection from while crosses later, taken crossing Wuchang ultra-deep ventilation shaft Wuhan subway as an example, influence by building studied. sphere influence, closer enclosure...