Dongsoo Han

In this paper we present a novel fast strain-limiting method that allows cloth to preserve its surface area efficiently. By preserving triangle area rather than edge length as other approaches do, this method does not remove the degrees of freedom of triangles and does not suffer from locking. Borrowing ideas from fluid simulation, we define pressures in each triangle and solve the global linear equation which shows a faster convergence over prior approaches which use Gauss-Seidel-like… Mehr zeigen

In this paper we present a novel fast strain-limiting method that allows cloth to preserve its surface area efficiently. By preserving triangle area rather than edge length as other approaches do, this method does not remove the degrees of freedom of triangles and does not suffer from locking. Borrowing ideas from fluid simulation, we define pressures in each triangle and solve the global linear equation which shows a faster convergence over prior approaches which use Gauss-Seidel-like iterations. The linear equation is easy to build by using edge and normal vectors and can be solved using Conjugate Gradient solver with regularization which not only helps the solver converge fast but also allows users to have a control over the stretchiness of cloth materials. Our area preserving strain limiting (APSL) can be also used as stand-alone cloth solver with linear bending springs. Weniger anzeigen

Implicit integration is a standard for stiff spring-based cloth simulation because of its stability. However constraints are useful to simulate various physical behaviors such as contact collisions or interaction with rigid bodies. Modified Conjugate Gradient (MCG) could support constraints as a part of implicit integration but constraints could not be added or removed during integration and they were limited to vertex nodes. Normally, a contact constraint has one or two frictional constraints… Mehr zeigen

Implicit integration is a standard for stiff spring-based cloth simulation because of its stability. However constraints are useful to simulate various physical behaviors such as contact collisions or interaction with rigid bodies. Modified Conjugate Gradient (MCG) could support constraints as a part of implicit integration but constraints could not be added or removed during integration and they were limited to vertex nodes. Normally, a contact constraint has one or two frictional constraints and act inside of triangle or edge rather than vertex node. Also its inequality property makes it harder to be included in MCG. For this reason, constraints are typically applied after implicit integration as a separate step or replaced with springs. In this paper, we propose a novel method to interleave various constraints with stiff springs so that we can take advantages from both sides. Also our Jacobian-free and matrix-free implicit integration allows us to use various nonlinear forces such as pressure or none vertex-centered forces. Interleaving collision constraints into integration step can eliminate unpleasant local deformation. Weniger anzeigen

This paper proposes a method to simulate inextensible hair strands using tridiagonal matrix formulation in which distance constraints are formulated as a linear system. The proposed method avoids constructing a full matrix explicitly. Instead, it takes advantage of the chain topology and serial indexing to formulate symmetric tridiagonal matrix. Furthermore, we use a linear distance constraint so that the constraint gradient can be easily formulated. With this matrix-free formulation, memory… Mehr zeigen

This paper proposes a method to simulate inextensible hair strands using tridiagonal matrix formulation in which distance constraints are formulated as a linear system. The proposed method avoids constructing a full matrix explicitly. Instead, it takes advantage of the chain topology and serial indexing to formulate symmetric tridiagonal matrix. Furthermore, we use a linear distance constraint so that the constraint gradient can be easily formulated. With this matrix-free formulation, memory usage can be extremely lowered. Since the formulated matrix is diagonally dominant, we can solve it by an efficient direct solver. Comparing error (i.e., stretch of constraints) of the proposed constraint solver to ones of the position-based solver with different number of iterations, we show that error of the proposed method is much smaller than those of position-based solver. Also the simulation result shows mush less numerical damping compared to Dynamic Follow-The-Leader method. By implementing in GPU, we demonstrate that our proposed method is simple and efficient. Weniger anzeigen

Hair can be a prominent feature of characters in real-time games. In this paper, we propose hair simulation with
efficient preservation of various hair styles. Bending and twisting effects are crucial to simulate curly or wavy hair.
We propose local and global shape constraints and parallel methods to update local and global transforms to find
goal positions. All three methods show good visual quality and take only a small fraction of rendering time. This
simulation runs on the GPU… Mehr zeigen

Hair can be a prominent feature of characters in real-time games. In this paper, we propose hair simulation with
efficient preservation of various hair styles. Bending and twisting effects are crucial to simulate curly or wavy hair.
We propose local and global shape constraints and parallel methods to update local and global transforms to find
goal positions. All three methods show good visual quality and take only a small fraction of rendering time. This
simulation runs on the GPU and works smoothly as a part of rendering pipeline. Simulating around 20,000 strands
composed of total 0.22 million vertices takes less than 1 millisecond. Simulation parameters such as stiffness or
number of iterations for shape constraints can be manipulated by users interactively. It helps designers choose
the right parameters for various hair styles and conditions. Also the simulation can handle various situations
interactively. Weniger anzeigen

This paper seeks to implement fluid simulation in tetrahedral mesh using hybrid method. Since Jos Stam introduced Stable Fluid in computer graphics community, FDM (finite difference method) in regular grid (hexahedral mesh) became widely used. The drawback of FDM in regular mesh has to do with the fact that treating curved boundary presents a challenge. To handle irregular boundary, tetrahedral mesh was used as a simulation domain. For advection, modified PIC/FLIP method was used to take… Mehr zeigen

This paper seeks to implement fluid simulation in tetrahedral mesh using hybrid method. Since Jos Stam introduced Stable Fluid in computer graphics community, FDM (finite difference method) in regular grid (hexahedral mesh) became widely used. The drawback of FDM in regular mesh has to do with the fact that treating curved boundary presents a challenge. To handle irregular boundary, tetrahedral mesh was used as a simulation domain. For advection, modified PIC/FLIP method was used to take advantage of low numerical dissipation. Unlike the traditional linear interpolation in PIC/FLIP, SPH kernel function based velocity interpolation method was experimented. Weniger anzeigen