No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Fast Cloth Animation on Walking Avatars
Computer Graphics Forum
Abstract
This paper describes a fast technique for animating clothing on walking humans. It exploits a mass-spring cloth model but applies a new velocity directional modification approach to overcome its super-elasticity. The algorithm for cloth-body collision detection and response is based on image-space interference tests, unlike the existing ones that use object-space checks. The modern workstations' graphics hardware is used not only to compute the depth maps of the body but also to interpolate the body normal vectors and velocities of each vertex. As a result the approach is very fast and makes it possible to produce animation at a rate of three to four frames per second.
... We can split the different animation approaches based on their goal: performance or realism. On one hand, Physically Based Simulation (PBS) [6,18,26,27,33,36,39] strategies discretize the space and time to apply basic physics laws. The realism obtained is closely related to how fine-grained is the discretization. ...
... The literature on this regard is exhaustive and mainly addresses the efficiency and robustness of the methodology. This is done through simplifications and specialization on constrained scenarios [6,26,27,36]. As another option, authors propose energy-based optimization approaches for an increase in stability and generalization to additional soft-bodies [18]. ...
We present a methodology to automatically obtain Pose Space Deformation (PSD) basis for rigged garments through deep learning. Classical approaches rely on Physically Based Simulations (PBS) to animate clothes. These are general solutions that, given a sufficiently fine-grained discretization of space and time, can achieve highly realistic results. However, they are computationally expensive and any scene modification prompts the need of re-simulation. Linear Blend Skinning (LBS) with PSD offers a lightweight alternative to PBS, though, it needs huge volumes of data to learn proper PSD. We propose using deep learning, formulated as an implicit PBS, to unsupervisedly learn realistic cloth Pose Space Deformations in a constrained scenario: dressed humans. Furthermore, we show it is possible to train these models in an amount of time comparable to a PBS of a few sequences. To the best of our knowledge, we are the first to propose a neural simulator for cloth. While deep-based approaches in the domain are becoming a trend, these are data-hungry models. Moreover, authors often propose complex formulations to better learn wrinkles from PBS data. Dependency from data makes these solutions scalability lower, while their formulation hinders its applicability and compatibility. By proposing an unsupervised methodology to learn PSD for LBS models (3D animation standard), we overcome both of these drawbacks. Results obtained show cloth-consistency in the animated garments and meaningful pose-dependant folds and wrinkles.
... Depending on the application we find two main classical computer graphics approaches. On the one hand, Physically Based Simulation (PBS) [6,19,25,26,31,32,35] approaches are able to obtain highly realistic cloth dynamics at the expense of a huge computational cost. On the other hand, Linear Blend Skinning (LBS) [14,15,17,22,33,34] and Pose Space Deformation (PSD) [3,4,18,20] models are suitable for environments with limited computational resources or realtime performance demand. ...
... Obtaining realistic cloth behaviour is possible through PBS (Physically Based Simulation), commonly through the well known mass-spring model. Literature on the topic is extensive, focused on improving the efficiency and stability of the simulation by simplifying and/or specializing on specific setups [6,25,26,32], or proposing new energy-based algorithms to enhance robustness, realism and generalization to other soft bodies [19]. Other works propose leveraging the parallel computational capabilities of modern GPUs [31,35]. ...
We present a novel approach to the garment animation problem through deep learning. Previous approaches propose learning a single model for one or few garment types, or alternatively, extend a human body model to represent multiple garment types. These works are not able to generalize to arbitrarily complex outfits we commonly find in real life. Our proposed methodology is able to work with any topology, complexity and multiple layers of cloth. Because of this, it is also able to generalize to completely unseen outfits with complex details. We design our model such that it can be efficiently deployed on portable devices and achieve real-time performance. Finally, we present an approach for unsupervised learning.
... Depending on the application we find two main classical computer graphics approaches. On the one hand, Physically Based Simulation (PBS) [6,19,25,26,31,32,35] approaches are able to obtain highly realistic cloth dynamics at the expense of a huge computational cost. On the other hand, Linear Blend Skinning (LBS) [14,15,17,22,33,34] and Pose Space Deformation (PSD) [3,4,18,20] models are suitable for environments with limited computational resources or realtime performance demand. ...
... Obtaining realistic cloth behaviour is possible through PBS (Physically Based Simulation), commonly through the well known mass-spring model. Literature on the topic is extensive, focused on improving the efficiency and stability of the simulation by simplifying and/or specializing on specific setups [6,25,26,32], or proposing new energy-based algorithms to enhance robustness, realism and generalization to other soft bodies [19]. Other works propose leveraging the parallel computational capabilities of modern GPUs [31,35]. ...
We present a novel approach to the garment animation problem through deep learning. Previous approaches propose learning a single model for one or few garment types, or alternatively, extend a human body model to represent multiple garment types. These works are not able to generalize to arbitrarily complex outfits we commonly find in real life. Our proposed methodology is able to work with any topology, complexity and multiple layers of cloth. Because of this, it is also able to generalize to completely unseen outfits with complex details. We design our model such that it can be efficiently deployed on portable devices and achieve real-time performance. Finally, we present an approach for unsupervised learning.
... Since the MSD model has a simple discrete structure, all kinds of operations including cut and suture in the surgery can be handled easily. This model has been widely used in facial animation (both static and dynamic, two dimensional and simplified three dimensional [23][24][25]), animating fire, clouds and water [40], animation of artificial animals [41,42], cloth draping [43,44], garment animation [45][46][47][48][49] and recently in surgical simulation [29,50]. To improve this model, a great deal of research has been carried out on various aspects-to refine the model adaptively [51], to update the Hooke's constants after refinement [30], to control the isotropy or anisotropy of the material [52], to improve its speed [28,29,53,54], to eliminate the super elasticity phenomena [55], and to handle post-buckling instability for stable but responsive simulation [49]. ...
... When deformation rate is greater than a critical deformation rate, the dynamic inverse procedure is applied to the two ends of the spring so that its deformation is exactly equal to the critical rate. Vassilev et al developed further Provot's method by applying a velocity directional modification approach to eliminate the super-elasticity phenomenon [48]. ...
... Therefore these methods often struggle to generate high-resolution garment animation efficiently. Some works attempt to reduce the computational cost [VSC01] by designing more efficient physical energy objective functions [GHF * 07, JGT17, LBK17] and position-based simulation [MHHR07]. Another line of work leverages CUDA acceleration and parallel algorithm design to efficiently simulate realistic garment animation [WWW22,Zel05] Figure 2: Overview of our method. ...
We address the 3D animation of loose‐fitting garments from a sequence of body motions. State‐of‐the‐art approaches treat all body joints as a whole to encode motion features, which usually gives rise to learned spurious correlations between garment vertices and irrelevant joints as shown in Fig. 1. To cope with the issue, we encode temporal motion features in a joint‐wise manner and learn an association matrix to map human joints only to most related garment regions by encouraging its sparsity. In this way, spurious correlations are mitigated and better performance is achieved. Furthermore, we devise the joint‐specific pose space deformation (PSD) to decompose the high‐dimensional displacements as the combination of dynamic details caused by individual joint poses. Extensive experiments show that our method outperforms previous works in most indicators. Moreover, garment animations are not interfered with by artifacts caused by spurious correlations, which further validates the effectiveness of our approach. The code is available at https://github.com/qiji77/JointNet.
... They produce robust collision detection, but their bottleneck is the update of the distance fields for deformable object. It can be alleviated for example using image-based approaches [VSC01]. Their memory cost is also an issue since the fields must be stored in a volumetric grid. ...
As animated films and series become more and more present in the mainstream entertainment, the artists’ needs are growing in term of fast and intuitive animation tools. Artists not only heavily rely on their imagination and skills to bring digital models to life; they also take inspiration from the physical world to better immerse viewers in their virtual environment.Many objects of our everyday surroundings exhibit elastic deformations when put in contact with others, e.g., a stress ball crushed by a hand, a pillow smashing a head during a pillow fight or a soft ball bouncing on a goal post. They most notably tend to squash inside the contact and to bulge outside of it. Such squashing and bulging effects are essential to communicate plausible deformation while capturing the physical behavior of soft materials in a variety of contexts, such as animated films. This type of deformation is, however, notoriously difficult and tedious to manually reproduce by computer graphics (CG) artists, and existing tools remain limited for artistic use.In practice, such deformations are thus generated through physically based simulation methods. However, due to their time-dependency, physical simulations must be run after the rigging and animation steps, preventing non-linear editing of the 3D scene. Moreover, artists also often resort to cartoonish deformation effects to better convey emotions and thoughts. Such exaggerated effects are difficult to achieve through physical simulations.The main contribution of this thesis is a novel purely geometric deformation framework that assists the artist by resolving local contacts between elastic objects and producing bulge effects in an art-directable way. To achieve a seamless integration within animation workflows, we designed our deformation tool to provide instant feedback to the artist while enabling non-linear editing thanks to a fully time-independent strategy. To produce plausible bulge effects, our method can also preserve the volume exactly, while artistic controls are also possible to explore more exaggerated behaviors. More specifically, starting from multiple meshes in intersection, our deformer first computes the parts of the surfaces remaining in contact, and then applies a procedural displacement controlled by a profile curve. Although our tool processes each frame independently, it achieves temporally continuous deformations with artistic control of the bulge through a small number of pseudo-stiffness parameters. The plausibility of the deformation is further enhanced by anisotropically spreading the volume-preserving bulge. An extension is also proposed to handle self-collisions between adjacent parts of the same object that often occur in character skinning animation.The result of this work is a robust, real-time deformer that can handle complex geometric configurations like a ball squashed by a hand, colliding lips, bending fingers, etc.
... Multiple viewpoints were placed around the body to produce depth maps against which particles were directly tested. This is similar to Vassilev et al. [23], who used whole body depth maps. The avatars were animated, but the pose was such that body parts did not occlude each other. ...
Many techniques facilitate real-time collision detection against complex models. These typically work by pre-computing information about the spatial distribution of geometry into a form that can be quickly queried. When models deform though, expensive pre-computations are impractical. We present radial fields: a variant of distance fields parameterised in cylindrical space, rather than Cartesian space. This 2D parameterisation significantly reduces the memory and computation requirements of the field, while introducing minimal overhead in collision detection tests. The interior of the mesh is defined implicitly for the entire domain. Importantly, it maps well to the hardware rasteriser of the GPU. Radial fields are much more application-specific than traditional distance fields. For these applications - such as collision detection with articulated characters - however, the benefits are substantial.
... The introduction of programmable GPUs allowed completely new algorithms. First, the depth buffer was used to determine overlapping primitives in conjunction with a color buffer for the collision response [VSC01], while the stencil buffer was used in CInDeR [KP03] to count ray-face intersections resulting from ray casts. CULLIDE [GRLM03] and its successor Quick-CULLIDE [GLM05] only used the image space for filtering potential collision candidates to reduce the errors caused by the 2.5D projection in image space. ...
We define a novel geometric predicate and a class of objects that enables us to prove a linear bound on the number of intersecting polygon pairs for colliding 3D objects in that class. Our predicate is relevant both in theory and in practice: it is easy to check and it needs to consider only the geometric properties of the individual objects – it does not depend on the configuration of a given pair of objects. In addition, it characterizes a practically relevant class of objects: we checked our predicate on a large database of real-world 3D objects and the results show that it holds for all but the most pathological ones. Our proof is constructive in that it is the basis for a novel collision detection algorithm that realizes this linear complexity also in practice. Additionally, we present a parallelization of this algorithm with a worst-case running time that is independent of the number of polygons. Our algorithm is very well suited not only for rigid but also for deformable and even topology-changing objects, because it does not require any complex data structures or pre-processing. We have implemented our algorithm on the GPU and the results show that it is able to find in real-time all colliding polygons for pairs of deformable objects consisting of more than 200k triangles, including self-collisions. © 2017 The Author(s) Computer Graphics Forum © 2017 The Eurographics Association and John Wiley & Sons Ltd. Published by John Wiley & Sons Ltd.
... The idea of this approach is to use non-uniform weights, which are small in the center of the grid and gradually increase as moving away from the center and reach their maximum at the boundary curves. In this way the addends in functional (1), which correspond to edges that are closer to the boundary curves, will have greater influence in the sum and the edges will follow the shape of the boundary curves (2). ...
This paper presents an approach for GPU acceleration of planar grid generation with a rectangular topology. To accelerate computing analyses, graphics units (GPU) are widely used as a low-cost, low-memory algorithms and high-performance computing platforms. The algorithm proposed in this paper is executed over a graphics processor and its main purpose is fast generation of large grids for cloth simulation with less memory usage. Резюме: Докладът представя подход за създаване на повърхнинни решетки с правоъгълна топология посредством програми, предназначени за паралелно програмиране. С цел увеличаването на изчислителната мощ се използват програми, изпълнявани върху графичен процесор. Основните причини са широкото им разпространение, ниската цена и най-вече високата изчислителна мощ. Алгоритъмът, представен в този доклад, се изпълнява именно върху графичен процесор и основните му цели са малко заемана памет и бързо генериране на големи по размери решетки за целите на виртуалните симулации на плат. Ключови думи: Виртуални Симулации на Плат, Създаване на Решетки, Графично ускорение INTRODUCTION In the recent years many developers in the field of Computer Graphics work on algorithms executing over large amount of data. One of these algorithms is generation of large grids used in cloth simulation. It requires a rectangular topology grid to be generated on a cloth piece, so that the cloth simulation of the masses and springs can run [1] [2] [3]. One of the most precise solutions is to use direct optimization method [4], but as shown in the next sections, it is not very fast for large grids. One of the newest research areas within numerical analysis is the parallel mesh generation. It is union between two scientific computing disciplines: computational geometry and parallel computing. Parallel mesh generation methods are using multiple processors or threads. One of the challenges in parallel mesh generation is to develop parallel meshing algorithm that generates meshes by using graphics units. The approach presented in this paper implements the improved length functional with non-uniform weights [5] and is executed over GPU. Detailed description of the solution is presented in the third section. The rest of the paper is organized as follows. The next section reviews previous work on grid generation. Section 3 describes in details the proposed approach. The last section concludes the paper and gives ideas for future work.
... Vassilev et al. [30] ...
Using an explicit integration method in physically based animations has many advantages including conceptual and computational simplicity, however, it re- quires small time steps to ensure low numerical instability. Simulations with large numbers of individually interacting components such as cloth, hair, and fluid models, are limited by the sections of particles most susceptible to error. This results in the need for smaller time steps than required for the majority of the system. These sections can be diverse and dynamic, quickly changing in size and location based on forces in the system. Identifying and handling these trou- blesome sections could allow for a larger time step to be selected, while preventing a breakdown in the simulation.
This thesis presents Smarticles (smart particles), a method of individually de- tecting particles exhibiting signs of instability and stabilizing them with minimal adverse effects to visual accuracy. As a result, higher levels of error introduced from large time steps can be tolerated with minimal overhead. Two separate approaches to Smarticles were implemented. They attempt to find oscillating particles by analyzing a particle’s (1) past behavior and (2) behavior with re- spect to its neighbors along a strand. Both versions of Smarticles attempt to correct unstable particles using velocity dampening. Smarticles was applied to a two dimensional hair simulation modeled as a continuum using smooth particle hydrodynamic. Hair strands are formed by linking particles together using one of two methods: position based dynamics or mass-spring forces.
Both versions of Smarticles, as well as a control of normal particles, were directly compared and evaluated based on stability and visual fluidity. Hair particles were exposed to various forms of external forces under increasing time step lengths. Testing showed that both versions of Smarticles working together allowed an average increase of 18.62% in the time step length for hair linked with position based dynamics. In addition, Smarticles was able to significantly reduce visible instability at even larger time steps. While these results suggest Smarticles is successful, the method used to correct particle instability may jeopardize other important aspects of the simulation. A more accurate correction method would likely need to be developed to make Smarticles an advantageous method.
... Most of the hardware-assisted image-space collision detection methods make use of OpenGL buffers such as depth buffer, stencil buffer, and color buffer (Baciu et al. 1999, Vassilev et al. 2001. But these methods require data read-back, which is often time-consuming due to the asymmetric accelerated graphics port buses in common graphic cards. ...
Collision detection plays a critical role in real-time applications such as game, simulation, and virtual reality. Collision avoidance is important in robotics path planning. Industrial safety, especially in construction and building, has a close linkage with the concept of contact avoidance. This chapter is interested in the investigation of collision detection problem using hardware graphics acceleration. Axis aligned bounding boxes (AABB) technique will be applied also for fast collision detection.
... Most of the hardware-assisted image-space collision detection methods make use of OpenGL buffers such as depth buffer, stencil buffer, and color buffer (Baciu et al. 1999, Vassilev et al. 2001. But these methods require data read-back, which is often time-consuming due to the asymmetric accelerated graphics port buses in common graphic cards. ...
... In this way the addends in functional (2), which correspond to edges that are closer to the boundary curves, will have greater influence in the sum and the edges will follow the shape of the boundary curves. (1) and (2). The coefficient k can be set to values greater or smaller than one to account for the different aspects in the physical domain. ...
This paper presents an approach for 3D grid generation with a rectangular topology. It is an improved direct optimization method. The idea of the paper is to provide a fast algorithm for mesh generation to do a more realistic 3D terrain models. Results and conclusions are presented at the end of the paper.
... [4] have used a hybrid explicit/implicit integration algorithm to animate real-time clothes which integrated as a Voxel-based collision detection algorithm but it seems to be limited by the maximum number of polygons that can animate in real-time. [18] proposed to use the z-buffer for collision detection to generate depth and normal maps. Computation time of their collision detection does not depend on the complexity of the body. ...
This paper presents the problem of performing collision detection between cloth and the solid object in real-time simulation. In order to achieve the fast collision detection and to help in collision computation, we propose a self-created Axis-Aligned Bounding Box (AABB) hierarchy. These AABBs can help in reducing the time needed to compute collision resolution.
... L'identification d'objets en collision peut être réalisée avec d'autres stratégies [29]. Certains préfèrent partitionner la scène en grilles régulières au lieu d'utiliser des boîtes englobantes [23,122]. Ou encore d'autres travaux ont été effectués sur l'emploi du matériel graphique dans la détection des collisions [137]. ...
Présidente de jury : Mme Marie-Paule Cani Rapporteur : M. Jean-Michel Dischler Rapporteur : M. Serge Miguet
... In the case of low neckline shirts it produces folded grids, as shown in Fig. 4. The idea of our approach is to compensate for this by using non-uniform weights, which are small in the centre of the grid and gradually increase as moving away from the centre and reach their maximum at the boundary curves. In this way the addends in functional (2), which correspond to edges that are closer to the boundary curves, will have greater influence in the sum and the edges will follow the shape of the boundary curves. ...
This paper presents two approaches for planar grid generation with a rectangular topology. One of them is algebraic and it is based on equality of quad areas and the other one is an improved direct optimisation method. The idea of the paper is to provide a fast algorithm for mesh generation to do a more realistic simulation of masses and springs. Results and conclusions are presented at the end of the paper.
... Its main applications are computer games, garment design and electronic commerce. Most of the mass-spring based approaches [1,2] require a rectangular topology grid to be generated on a cloth piece, so that the physical simulation of the masses and springs can run. One of the easiest solutions is to use an algebraic method [3], but as shown in the next sections, it does not always produce good results. ...
This paper presents two approaches for planar grid generation. One of them is algebraic and it is based on equality of quad areas and the other one is an improved direct optimization method. The idea of the paper is to provide a fast algorithm for mesh generation to do a more realistic simulation of mass-spring models. Results and conclusions are presented at the end of the paper.
... Bei animierten Menschmodellen wird dazu eine Kollisionskörperhierarchie aufgebaut. Weitere Möglichkeiten wären die Voxelisierung des Modells[FSG03] oder die Nutzung von bildbasierten Verfahren[RNS05,RN11,VSC01] ...
In diesem Paper wird eine Erweiterung des Position-Based-Dynamics-Verfahrens vorgestellt, mit deren Hilfe die Kollisionsbehandlung für Kleidung an animierten Menschmodellen deutlich vereinfacht wird. Nicht oder zu spät erkannte Kollisionen können, je nach Verfahren, besonders bei schnellen Bewegungen zu falsch dargestellter Kleidung führen. Aus diesem Grund wird das Position-Based-Dynamics-Verfahren um ein Animationsconstraint erweitert. Dieses ermöglicht einen frei definierbaren Anpassungsgrad der Kleidung an die Bewegungen des Menschmodells, ohne auf eine physikalisch korrekte Kleidungsdarstellung zu verzichten.
... Another way to speed up deformable collision detection is by exploiting spatial coherency [82]. Special attention has been devoted to the problem of collision detection for moving cloth [130,107,129,97,28], which is especially challenging due to frequent self-collisions. A rather brute force GPU algorithm has been proposed by Choi at el. [25]. ...
Due to the highly nonlinear behavior of clothing, modelling fine-scale garment deformation on arbitrary meshes under varied conditions within a unified network poses a significant challenge. Existing methods often compromise on either model generalization, deformation quality, or runtime speed, making them less suitable for real-world applications. To address it, we propose to incorporate multi-source graph construction and pooling to achieve a novel graph learning scheme. We first introduce methods for extracting cues from different deformation correlations and transform the garment mesh into a comprehensive graph enriched with deformation-related information. To enhance the learning capability and generalizability of the model, we present structure-preserving pooling and unpooling strategies for the mesh deformation task, thereby improving information propagation across the mesh and enhancing the realism of deformation. Lastly, we conduct an attribution analysis and visualize the contribution of various vertices in the graph to the output, providing insights into the deformation behavior. The experimental results demonstrate superior performance against state-of-the-art methods.
We present a methodology to automatically obtain Pose Space Deformation (PSD) basis for rigged garments through deep learning. Classical approaches rely on Physically Based Simulations (PBS) to animate clothes. These are general solutions that, given a sufficiently fine-grained discretization of space and time, can achieve highly realistic results. However, they are computationally expensive and any scene modification prompts the need of re-simulation. Linear Blend Skinning (LBS) with PSD offers a lightweight alternative to PBS, though, it needs huge volumes of data to learn proper PSD. We propose using deep learning, formulated as an implicit PBS, to un-supervisedly learn realistic cloth Pose Space Deformations in a constrained scenario: dressed humans. Furthermore, we show it is possible to train these models in an amount of time comparable to a PBS of a few sequences. To the best of our knowledge, we are the first to propose a neural simulator for cloth. While deep-based approaches in the domain are becoming a trend, these are data-hungry models. Moreover, authors often propose complex formulations to better learn wrinkles from PBS data. Supervised learning leads to physically inconsistent predictions that require collision solving to be used. Also, dependency on PBS data limits the scalability of these solutions, while their formulation hinders its applicability and compatibility. By proposing an unsupervised methodology to learn PSD for LBS models (3D animation standard), we overcome both of these drawbacks. Results obtained show cloth-consistency in the animated garments and meaningful pose-dependant folds and wrinkles. Our solution is extremely efficient, handles multiple layers of cloth, allows unsupervised outfit resizing and can be easily applied to any custom 3D avatar.
We present a CPU-based real-time cloth animation method for dressing virtual humans of various shapes and poses. Our approach formulates the clothing deformation as a high-dimensional function of body shape parameters and pose parameters. In order to accelerate the computation, our formulation factorizes the clothing deformation into two independent components: the deformation introduced by body pose variation (Clothing Pose Model) and the deformation from body shape variation (Clothing Shape Model). Furthermore, we sample and cluster the poses spanning the entire pose space and use those clusters to efficiently calculate the anchoring points. We also introduce a sensitivity-based distance measurement to both find nearby anchoring points and evaluate their contributions to the final animation. Given a query shape and pose of the virtual agent, we synthesize the resulting clothing deformation by blending the Taylor expansion results of nearby anchoring points. Compared to previous methods, our approach is general and able to add the shape dimension to any clothing pose model. Furthermore, we can animate clothing represented with tens of thousands of vertices at 50+ FPS on a CPU. We also conduct a user evaluation and show that our method can improve a user's perception of dressed virtual agents in an immersive virtual environment (IVE) compared to a realtime linear blend skinning method.
In the process of simulating and modeling real objects, the phenomenon of objects penetrating each other may occur in the model, which is unrealistic and then the research of collision detection (CD) is generated. As a bottleneck of virtual environment simulation, researchers have conducted in‐depth research on CD, especially the CD of deformable objects. In this paper, we briefly review the related research on CD of deformable objects regarding relevant literature. First, we briefly introduce previous reviews of CD. Second, we review the popular research methods and limitations of CD between deformable objects. Third, we review the popular research methods and limitations of self‐collision detection in deformable objects. Finally, we discuss future directions of development. This review can be used as a reference for the application of CD in all directions.
This paper proposes an improved mass-spring elastic model of cloth by applying a shape matching technique to the original Hookean springs. As a result the springs do not obey to the Hook's law anymore, which allows stable simulation of cloth with larger time steps, with the simplicity and efficiency of the mass-spring system preserved. The proposed model was implemented in a system for dressing garments on virtual human bodies. Test results are given at the end of the paper, which show the advantages of the new approach, compared to the original mass-spring model.
In the paper, the mass-spring model was used as the basic cloth simulation model to simulate the skirt. After modeling, different forces were applied on the particles and generated differential equations. To solve the equations, different kinds of integration methods was used in the paper for comparisons of getting the best results. To correct the super-elastic phenomenon usually occurring in cloth simulation, the improved mass-constrained method was also adopted. The paper mainly shows the process of generating a skirt simulation model with obviously realistic wrinkles and folds fitting with an elliptic-shaping waist model.
Purpose
The purpose of this paper is to conduct a survey on research in fabric and cloth simulation using mass spring model. Also in this paper some of the common methods in process of fabric simulation in mass spring model are discussed and compared.
Design/methodology/approach
This paper reviews and compares presented mesh types in mass spring model, forces applied on model, super elastic effect and ways to settle the super elasticity problem, numerical integration methods for solving equations, collision detection and its response. Some of common methods in fabric simulation are compared to each other. And by using examples of fabric simulation, advantages and limitations of each technique are mentioned.
Findings
Mass spring method is a fast and flexible technique with high ability to simulate fabric behavior in real time with different environmental conditions. Mass spring model has more accuracy than geometrical models and also it is faster than other physical modeling.
Originality/value
In the edge of digital, fabric simulation technology has been considered into many fields. 3D fabric simulation is complex and its implementation requires knowledge in different fields such as textile engineering, computer engineering and mechanical engineering. Several methods have been presented for fabric simulation such as physical and geometrical models. Mass spring model, the typical physically based method, is one of the methods for fabric simulation which widely considered by researchers.
We introduce a new cloth simulation algorithm that uses a special data of nodes-springs-elements (NSE) order that able to achieve to a realistic cloth simulation. We develop ordering procedures on data of spring-mass system that represents physics and geometrics of a cloth as in reality. In our experiments we developed a rectangle cloth model which is initially at a horizontal position with one fixed point and it is allowed to drape by its own weight. Our simulation result is not only able to achieve a plausible cloth drapes but also reduce overstretching when large time step is used. This paper aims to demonstrate the reliability of our approach to eliminate overstretching and ability to control bending properties during simulation.
A cloth simulation represents the behavior of cloth objects such as flag tablecloth or even garments has application in clothing animation for games and virtual shops. Elastically deformable models have widely used to provide realistic and efficient simulation however problem of overstretching is encountered. We introduce a new cloth simulation algorithm that replaces iterative distance constraint enforcement steps with non-iterative ones for preventing over stretching in a spring-mass system for cloth modeling. Our method is based on a simple position correction procedure applied at one end of a spring. In our experiments we developed a rectangle cloth model which is initially at a horizontal position with one point is fixed and it is allowed to drape by its own weight. Our simulation is able to achieve a plausible cloth drapes as in reality. This paper aims to demonstrate the reliability of our approach to overcome overstretches while decreasing the computational cost of the constraint enforcement process due to an iterative procedure that is eliminated.
Im Wintersemester 2011/2012 hat am Lehrstuhl 7 (Graphische Systeme) der Fakultät
Informatik der TU Dortmund in Kooperation mit dem Institut für Spanende Fertigung die
Projektgruppe ChipSim, mit der Aufgabe, ein Simulations- und Visualisierungssystem zu
entwerfen und zu implementieren, um die Spanbildung untersuchen zu können, begonnen.
Das entwickelte System soll plastoelastische Objektinteraktionen realistisch darstellen.
Dabei ist es notwendig, dass die Kinematik der simulierten Objekte durch bestimmte
externe und interne physikalische Kräfte sowie durch Kollisionen von Schneide und
Werkstück beeinflusst wird. Als externe Kraft kann z.B. die Gravitationskraft, als interne
Kräfte bzw. Eigenschaften z.B. der Druck und die Viskosität berücksichtigt werden.
When realistic motion is the objective, generating motion can be based on computations that use basic physical laws, usually consisting of forces applied to masses that produce accelerations. This chapter starts by reviewing the basic forces that are important to computer animation: gravity, springs, dampers, friction, and force fields (e.g., wind). Once basic physics is reviewed, some of the main applications in which physics is used in computer animation are presented, such as spring meshes, particle systems, rigid body simulation, and cloth. The chapter concludes with a discussion of enforcing soft and hard constraints.
In the current chapter, the authors intend to present the strengths and the weaknesses from the standpoint of graphic computing, managing and productive computing, and the human factors that prevent boosting on-line sales. They analyze the reality of several businesses or industries, which are gathered under the word "group" and belong to the textile sector with high profits because of the high volume of invoicing, a hundred-year-old tradition, with important customers of the European, American, and Asian international fashion, etc. in a northern Italian region known as Lombardy. Paradoxically, the textile sector in that region, like others in Europe, is one of the environments where computing and the breakthroughs and the advantages in new technologies are difficult to introduce these days. The human factor is the main cause because of organizational structures of a vertical and/or family type.
Whether collision detection tests interference occurred between objects, it is the key technique in virtual reality. According to different objects, we should choose different ways. Author introduces the collision detection algorithm and research status in detail, analyses the various algorithm and offers the basis theory for the deeper development of collision detection in virtual reality.
The chapter begins by discussing the modeling and simulation techniques for garments. It reviews the three approaches for modeling garments: geometry based, physically based and hybrid. Computer graphics techniques and collision detection methods are then described for interactive garment simulation. To render garments realistically, they are discretized as polygons, with patterns repeated over the whole surface of the garments, and a bidirectional reflectance distribution function is then used for computing the light interaction with the surface of the garments. The future development of modeling and simulation of garment materials is discussed at the end of the chapter.
With the development of computer graphics technologies, modeling and simulation of cloth animation becomes a common subject in 3D game and animated film related field, and presents the booming trend. The main objective of cloth animation is to obtain rich details of cloth movement and realistic animation. It can be widely used in games, animation and virtual reality, and many other fields. For its complicated nonlinear, anisotropic elastic behavior of the cloth, natural and realistic clothing wrinkles and shapes are difficult to create. The researches on cloth animation based on physical simulation are reviewed and the basic theory, method and framework are summarized. For the widely used methods in cloth animation, the essential process such as modeling, solving dynamics equations, collision detection and response are introduced. With analyzing the related literatures, the advantages, disadvantages and interrelationship of the existing modeling and simulation methods are summarized. Finally, from the demand of realistic performance and real-time interaction, future directions in modeling and simulation of cloth animation are discussed.
Driven by demand from the entertainment industry for better and more realistic animation, technology continues to evolve and improve. The algorithms and techniques behind this technology are the foundation of this comprehensive book, which is written to teach you the fundamentals of animation programming. In this third edition, the most current techniques are covered along with the theory and high-level computation that have earned the book a reputation as the best technically-oriented animation resource. Key topics such as fluids, hair, and crowd animation have been expanded, and extensive new coverage of clothes and cloth has been added. New material on simulation provides a more diverse look at this important area and more example animations and chapter projects and exercises are included. Additionally, spline coverage has been expanded and new video compression and formats (e.g., iTunes) are covered. * Includes companion site with contemporary animation examples drawn from research and entertainment, sample animations, and example code * Describes the key mathematical and algorithmic foundations of animation that provide you with a deep understanding and control of technique * Expanded and new coverage of key topics including: fluids and clouds, cloth and clothes, hair, and crowd animation * Explains the algorithms used for path following, hierarchical kinematic modelling, rigid body dynamics, flocking behaviour, particle systems, collision detection, and more.
We present a non-iterative approach to visually plausible rendering of multi-layered clothes, aimed mainly at human figures dressed in multiple garments. The problems we address are cloth interference and z-fighting. Our approach is based on numbering layers of clothing and then using recursive ray tracing to keep them in order. Instead of developing a new physically based technique, such as collision detection/response schemes or particle systems, we develop a visualization technique to complement existing physically based methods, greatly reducing the required precision and, thus, the computational cost.
This chapter gives the basic method of representing and modeling a freeform product. Not only human-centered products but also any other freeform objects can be represented and modeled by a non-manifold data structure and its related operators introduced here. Based on this, two approaches to design human-centered products are developed. A constructive method is introduced for designing products whose spatial relationship with human bodies is explicitly defined during the design procedure. A surface subdivision method is then employed to refine constructive designs into final surfaces. The other method is more user-friendly, which allows users to interactively draw styling design curves of tight-fit products on the surfaces of digital human bodies. Final products are generated by trimming the surfaces of human bodies with these styling curves.
In the clothing industry, compression garments are increasingly being used to mold and confine the malleable shapes of human bodies. A garment design may require variations in pressure ranging from normal to larger values through increasing strains in specific areas for particular applications. Therefore, compression garments need to be customized because body shapes differ and different body shapes require different strain distributions. 3D body shapes can be obtained by any popular 3D data-acquisition means (for example, a human body laser scanner). 2D patterns which are fabricated into the 3D shape bring about the desired strain distribution by the variation of their shapes between 2D and 3D. At present, the 2D pattern design and garment-to-body fit are accomplished by trial and error. This chapter aims at providing techniques to automate this design procedure.
The challenges of virtual garment simulation are numerous and have attracted research efforts for more than a decade. First dedicated to the realistic simulation of the mechanical behavior of cloth, virtual garment simulation soon evolved toward the simulation of virtual garments on synthetic characters. Although the computer graphics industry reaps the most obvious benefits from garment simulation on animated virtual characters, virtual prototyping of garment models is another major application field for the garment industry.
We propose a graphics processing unit-based approach to accelerate the radial view-based culling method for continuous self-collision detection of deformable surfaces. The deformable surfaces may have small round-shaped holes and ghost triangles are used to fill the holes. We identify the key processes of the radial view-based culling method, including triangle classification, traversal of bounding volume hierarchies and handling violated triangles (i.e., the triangles intersecting with ghost triangles). We propose efficient parallel processing techniques to perform these key processes on a programmable graphics unit. We have evaluated our proposed approach on several examples. Experimental results show that our approach significantly cuts down the cost of the key processes of the radial-based culling method, compared with the serial implementation on CPU.
To reduce the cost of correcting design errors, assemblies of mechanical parts are modeled using CAD systems and verified electronically before the designs are sent to manufacturing. Shaded images are insufficient for examining the internal structures of assemblies and for detecting interferences. Thus, designers must rely on expensive numerical techniques that compute geometric representations of cross-sections and of intersections of solids. The solid-clipping approach presented here bypasses these geometric calculations and offers realtime rendering of cross-sections and interferences for solids represented by their facetted boundaries. In its simplest form, the technique is supported by contemporary highend graphics workstations. Its variations, independently developed elsewhere, have already been demonstrated. Our implementation is based on the concept of a cutvolume interactively manipulated to remove obstructing portions of the assembly and reveal its internal structure. For clarity, faces of the cut-volume which intersect a single solid are hatched and shaded with the color of that solid. Interference areas between two or more solids are highlighted. Furthermore, to help users find the first occurrence of an interference along a search direction, we have developed an adaptive subdivision search based on a projective approach which guarantees a sufficient condition for object disjointness. The additional performance cost for solid-clipping and interference highlighting is comparable to the standard rendering cost. An efficient implementation of the disjointness test requires a minor extension of the graphics functions currently supported on commercial hardware.
We develop physically-based models of deformable curves, surfaces, and solids for use in computer graphics. Our deformable models are governed by the mechanical laws of continuous bodies whose shapes can change over time. These laws, expressed in the form of dynamic differential equations, unify the description of shape and motion. By solving the equations numerically we are able to create realistic animations involving the interaction of deformable models with various applied forces, ambient media, and impenetrable obstacles in a simulated physical world. We develop deformable models capable of perfectly elastic behavior and more general inelastic behavior, including viscoelasticity, plasticity, and fracture.
Animating the drape of different cloths must address complex physical behaviors. This particle approach uses optimizations that make it faster than earlier implementations and allow it to simulate behavior over time. The modeling system presented computes the full trajectories of particles and not just the final positions. This offers several important advantages. Since the full history of each particle is known, hysteresis effects can be modeled accurately. The Kawabata (1980) experimental data for different textiles can be input directly to the model. The effects of external forces, especially those produced by wind or moving solid bodies, can be modeled accurately. Despite this extra dimension of detail, our system computes final positions considerably faster than the times given by Breen, House and Wozny (1994). Our model can be easily extended to simulate the effects of manufacturing processes or interacting bodies. In particular, high stresses of the kind that occur in manufacturing can only be modeled if the full trajectory of each particle is known. We have implemented our model as a C++ class library. Particle systems are more flexible than approaches using continuum mechanics. Our system's fast computation times, mainly due to the numerical solution of ordinary differential equations, compare favorably to approaches using a finite-element method. Therefore, our approach might be an interesting alternative for other engineering problems currently solved by a finite-element method, for example, the computation of minimal surfaces, heavy membranes, vibrating membranes and population dynamics
This paper discusses the use of physics-based models for animating clothes on synthetic actors in motion. In our approach, cloth pieces are first designed with polygonal panels in two dimensions, and are then seamed and attached to the actor's body in three dimensions. After the clothes are created, physical properties are simulated and then clothes are animated according to the actor's motion in a physical environment. We describe the physical models we use and then address several problems we encountered. We examine how to constrain the elements of deformable objects which are either seamed together or attached to rigid moving objects. We also describe a new approach to the problem of handling collisions among the cloth elements themselves, or between a cloth element and a rigid object like the human body. Finally, we discuss how to reduce the number of parameters for improving the interface between the animator and the physics-based model.
We present an efficient algorithm for detecting self collisions, as well as some techniques for evaluating collision inside-outside orientation in a robust way. As presented in [VOL 94], we detect collisions using a hierarchical algorithm that takes advantage of curvature properties giving us full power of hierarchical algorithms for self-collision situations. Determining the collision orientation may become a complex problem dealing with complex collisions reulting from highly deformable surfaces. We use collision remnance and consistency correction for computing collision orientation in a robust way, for accurate collision response in simulations involving highly deformed and wrinkled surfaces. Keywords : Collision detection, self-collision, geometrical regularity, automatic hierarchisation, adjacency detection, surface orientation, consistency. 1. Introduction Mechanical simulation of soft surfaces usually require collision detection to be performed, for avoiding the simulated ob...
Interference detection is a useful technique, but it is also generally time‐consuming. In this paper, a new type of interference detection algorithm is proposed for real‐time interference detection. The algorithm first rasterizes the projection of the target objects and calculates the z ‐values, just as done by the z ‐buffer visible surface algorithm. For interference detection, all z ‐values and pointers to the corresponding faces of objects are saved in a z ‐list for each pixel. Sorting the z ‐list against the z ‐values allows the detection of overlapping objects in the z ‐direction at each pixel position and, thus, finds interfering faces by referring to the face pointers in the z ‐list.
The algorithm is simple and easy to implement. Its computational complexity is directly proportional to the number of polygons, and, in addition, standard graphics hardware can be used to greatly accelerate execution. Another advantage is that the algorithm can be applied to all ‘ray‐traceable’ objects, including algebraic surfaces, and procedurally defined objects; traditionally these were not suitable subjects for interference detection.
The algorithm is implemented on a graphics workstation using a standard graphics library. Interference detection at a practical interaction speed is achieved for complicated objects such as polyhedra with thousands of polygons. The algorithm can be used in two ways: for inexpensive interference detection, and as an efficient culling method for more precise collision/interference detection algorithms.
This article presents a method for collision handling applied to the semi-rigid mass-spring cloth model formerly described in [Pro95].
This method deals with the four main difficulties encountered in collision handling. The first is collision detection. The second is optimization of collision detection, which is otherwise excessively time consuming. The third is collision response. The fourth is conservation of collision consistency. The latter is discussed in detail, and related to cases of interfering multiple collisions. An original method for computation of collision response in this case of multiple collisions is presented, providing a robust conservation of collision consistency.
Results obtained with this approach, in the case of building real garments on a mannequin, are presented and validate our cloth model and collision handling method.
We present an efficient method for detecting collisions between complex solid objects. The method features a stable processing
time and low sensitivity to the complexity of contact between objects. The algorithm handles both concave and convex objects;
however, the best performance is achieved when at least one object is convex in the proximity of the collision zone (our techniques
check the required convexity property as a byproduct of the calculations). The method achieves real-time performance when
calculations are supported by the standard functionality of graphics hardware available on high-end workstations.
Object interactions are ubiquitous in interactive computer
graphics, 3D object motion simulations, virtual reality and robotics
applications. Most collision detection algorithms are based on
geometrical object space interference tests. Some algorithms have
employed an image space approach to the collision detection problem. We
demonstrate an image space collision detection process that allows
substantial computational savings during the image space interference
test. This approach makes efficient use of the graphics rendering
hardware for real time complex object interactions
In this survey, we present a contemporary overview of cloth modeling techniques. 19 modeling techniques are summarized and categorized by their main theoretical method: geometrical, physical, or hybrid. The techniques within each category do not follow well-defined patterns. We therefore generally report each work independently according to the chronology of publication. At the end of the discussion of all techniques, we summarize their features in a table. We conclude by speculating on future research directions that could optimize the agreement between the requirements of visual realism and physical accuracy. The recommendations for future work consider the different goals in textile engineering and computer graphics
The bottle-neck in most cloth simulation systems is that time steps must be small to avoid numerical instability. This paper describes a cloth simulation system that can stably take large time steps. The simulation system couples a new technique for enforcing constraints on individual cloth particles with an implicit integration method. The simulator models cloth as a triangular mesh, with internal cloth forces derived using a simple continuum formulation that supports modeling operations such as local anisotropic stretch or compression; a unified treatment of damping forces is included as well. The implicit integration method generates a large, unbanded sparse linear system at each time step which is solved using a modified conjugate gradient method that simultaneously enforces particles' constraints. The constraints are always maintained exactly, independent of the number of conjugate gradient iterations, which is typically small. The resulting simulation system is significantly faster than previous accounts of cloth simulation systems in the literature.
This paper describes a physically-based model for animating cloth objects, derived from elastically deformable models, and improved in order to take into account the non-elastic properties of woven fabrics. A cloth object is first approximated to a deformable surface composed of a network of masses and springs, the movement of which is evaluated using the numerical integration of the fundamental law of dynamics. We show that when a concentration of high stresses occurs in a small region of the surface, the local deformation becomes unrealistic compared to real deformations of textiles. With such an elastic model, the only solution to decrease these deformations has been so far to increase the stiffness of the deformed springs, but we show that it dramatically increases the cost of the algorithm. We present therefore a new method to adapt our model to the particularly stiff properties of textiles, inspired from dynamic inverse procedures.
Course note 12, Introduction to Computer Graphics
- N I Badler
- A Glassner
Badler N I, Glassner A S. 3D object modelling.
Course note 12, Introduction to Computer Graphics.
SIGGRAPH 1998; 1-14.
Interference detection through rasterization
- M Shinya
- M Forque
Shinya M and Forque M. Interference detection
through rasterization. Journal of Visualization and
Computer Animation 1991; 2:131-134.
Interactive inspection of solids: cross-section and interferences
- J Rossignac
- A Megahed
- B Schneider
Rossignac J, Megahed A, Schneider B O.
Interactive inspection of solids: cross-section and
interferences. Computer Graphics (SIGGRAPH
1992); 26(2): 353-360.
RECODE: an imagebased collision detection algorithm
- G Baciu
- W S Wong
- H Sun
Baciu G, Wong W S, Sun H. RECODE: an imagebased collision detection algorithm. The Journal of
Visualization and Computer Animation 1999; 10
(4): 181-192.
- D Terzopoulos
- J Platt
- Barr A Fleischer
Terzopoulos D, Platt J, Barr A and Fleischer K.
Elastically Deformable Models. Computer Graphics
(Proc. SIGGRAPH 1987); 21 (4): 205-214.
Predicting the drape of woven cloth using interacting particles
- D E Breen
- D House
- M J Wozhny
Breen D E, House D H and Wozhny M J. Predicting
the drape of woven cloth using interacting particles.
Computer Graphics (Proc. SIGGRAPH 1994);
28:23-34.