A nonparametric regression model for virtual humans generation

In this paper, we propose a novel nonparametric regression model to generate virtual humans from still images for the applications of next generation environments (NG). This model automatically synthesizes deformed shapes of characters by using kernel regression with elliptic radial basis functions (ERBFs) and locally weighted regression (LOESS). Kernel regression with ERBFs is used for representing the deformed character shapes and creating lively animated talking faces. For preserving patterns within the shapes, LOESS is applied to fit the details with local control. The results show that our method effectively simulates plausible movements for character animation, including body movement simulation, novel views synthesis, and expressive facial animation synchronized with input speech. Therefore, the proposed model is especially suitable for intelligent multimedia applications in virtual humans generation.     

Estimating body shape of dressed humans

The paper presents a method to estimate the detailed 3D body shape of a person even if heavy or loose clothing is worn. The approach is based on a space of human shapes, learned from a large database of registered body scans. Together with this database we use as input a 3D scan or model of the person wearing clothes and apply a fitting method, based on ICP (iterated closest point) registration and Laplacian mesh deformation. The statistical model of human body shapes enforces that the model stays within the space of human shapes. The method therefore allows us to compute the most likely shape and pose of the subject, even if it is heavily occluded or body parts are not visible. Several experiments demonstrate the applicability and accuracy of our approach to recover occluded or missing body parts from 3D laser scans.     

一种虚拟人追逐过程中的情绪博弈模型

追逐行为在人类社会生活中十分常见,模拟追逐过程中的情绪变化,有助于揭示情绪和运动的关系.在借鉴博弈论的基础上,研究了虚拟人围绕障碍物追逐的情绪博弈模型,依据追逐的剧情,拟定了虚拟人的情绪强度公式,定量描述了虚拟人的速度和体力的变化,提出了追逐过程中虚拟人情绪博弈策略.采用OpenSceneGraph图形库实现了虚拟人追逐动画,实验结果表明,该模型能很好地模拟追逐中虚拟人情绪的变化规律.     

基于Web服务的统一虚拟系统管理

为了解决虚拟化技术的发展过程中,虚拟化产品的多样性带来的虚拟系统管理的复杂性问题.DMIT提出的WBEM是实现企业系统资源统一管理的标准框架,规范了企业网络中受管资源的描述与使用.通过WBEM企业系统资源统一管理框架中的CIM资源建模方式对虚拟系统资源进行统一建模,并结合最新的基于Web服务的标准化CIM信息交换接口即WS-Management标准,实现了通用的虚拟系统管理原型,以解决各种异构虚拟化平台统一管理问题.     

Physical simulation of wet clothing for virtual humans

We present a technique that simulates wet garments for virtual humans with realistic folds and wrinkles. Our approach combines three new models to allow realistic simulation of wet garments: (1) a simplified saturation model that modifies the masses, (2) a nonlinear friction model derived from previously reported, real-world measurements, and (3) a wrinkle model based on imperfection sensitivity theory. In contrast to previous approaches to wet cloth, the proposed models are supported by the experimental results reported in the textile literature with parameters varying over the course of the simulation. As a result, the wet garment motions simulated by our method are comparable to that of real wet garments. Our approach recognizes the special, practical importance of contact models with human skin and provides a specific skin-cloth friction solution. We evaluate our approach by draping a rotating sphere and simulating a typical garment on a virtual human in the rain. Both of these examples are typical scenarios within computer graphics research.     

A unified model for abduction-based reasoning

In the last decade, abduction has been a very active research area. This has resulted in a variety of models mechanizing abduction, namely within a probabilistic or logical framework. Recently, a few abductive models have been proposed within a neural framework. Unfortunately, these neural/probablistic/logical-based models cannot address complex abduction problems. In this paper, we propose a new extended neural-based model to deal with abduction problems which could be monotonic, open, and incompatible     

Workspace zone differentiation and visualization for virtual humans

Human performance measures such as discomfort and joint displacement play an important role in product design. The virtual human Santos, a new generation of virtual humans developed at the University of Iowa, goes directly to the computer-aided design model to evaluate a design, saving time and money. This paper presents an optimization-based workspace zone differentiation and visualization. Around the workspace of virtual humans, a volume is discretized to small zones and the posture prediction on each central point of the zone will determine whether the points are outside the workspace as well as the values of different objective functions. Visualization of zone differentiation is accomplished by showing different colours based on values of human performance measures on points that are located inside the workspace. The proposed method can subsequently help ergonomic design. For example, in a vehicle's interior, the controls should not only lie inside the workspace, but also in the zone that encloses the most comfortable points. Using the palette of colours inside the workspace as a visual guide, a designer can obtain a reading of the discomfort level of product users.     

Workspace zone differentiation and visualization for virtual humans

Human performance measures such as discomfort and joint displacement play an important role in product design. The virtual human Santos?, a new generation of virtual humans developed at the University of Iowa, goes directly to the computer-aided design model to evaluate a design, saving time and money. This paper presents an optimization-based workspace zone differentiation and visualization. Around the workspace of virtual humans, a volume is discretized to small zones and the posture prediction on each central point of the zone will determine whether the points are outside the workspace as well as the values of different objective functions. Visualization of zone differentiation is accomplished by showing different colours based on values of human performance measures on points that are located inside the workspace. The proposed method can subsequently help ergonomic design. For example, in a vehicle's interior, the controls should not only lie inside the workspace, but also in the zone that encloses the most comfortable points. Using the palette of colours inside the workspace as a visual guide, a designer can obtain a reading of the discomfort level of product users.     

A unified neurofuzzy model for classification

This work proposes a unified neurofuzzy modelling scheme. To begin with, the initial fuzzy base construction method is based on fuzzy clustering utilising a Gaussian mixture model (GMM) combined with the analysis of covariance (ANOVA) decomposition in order to obtain more compact univariate and bivariate membership functions over the subspaces of the input features. The mean and covariance of the Gaussian membership functions are found by the expectation maximisation (EM) algorithm with the merit of revealing the underlying density distribution of system inputs. The resultant set of membership functions forms the basis of the generalised fuzzy model (GFM) inference engine. The model structure and parameters of this neurofuzzy model are identified via the supervised subspace orthogonal least square (OLS) learning. Finally, instead of providing deterministic class label as model output by convention, a logistic regression model is applied to present the classifier’s output, in which the sigmoid type of logistic transfer function scales the outputs of the neurofuzzy model to the class probability. Experimental validation results are presented to demonstrate the effectiveness of the proposed neurofuzzy modelling scheme.     

Motion control of virtual humans

The article surveys virtual humans and techniques to control the face and body. It also covers higher level interfaces for direct speech input and issues of real-time control     

An ontology of virtual humans

Most of the efforts concerning graphical representations of humans (Virtual Humans) have been focused on synthesizing geometry for static or animated shapes. The next step is to consider a human body not only as a 3D shape, but as an active semantic entity with features, functionalities, interaction skills, etc. We are currently working on an ontology-based approach to make Virtual Humans more active and understandable both for humans and machines. The ontology for Virtual Humans we are defining will provide the “semantic layer” required to reconstruct, stock, retrieve, reuse and share content and knowledge related to Virtual Humans.     

Reusable,lifelike virtual humans for mentoring and role-playing

Lifelike, interactive digital characters, serving as mentors and role-playing actors, have been shown to significantly improve learner motivation and retention. However, the cost of modeling such characters, authoring and editing their interactions, and delivering them over limited-bandwidth connections can be prohibitive. This paper describes a framework, authoring tools, and Web-based run-time environment that support the creation of training scenarios using digital virtual humans and other reusable 3D components. By conforming to the Humanoid Animation (H-Anim), Extensible 3D, and ADL Shareable Content Object Reference Model specifications, these 3D components are designed to promote reuse and interoperability at several levels. Recently, these software tools were used to develop prototype lessons in foreign language and cultural familiarization for use at the Defense Language Institute Foreign Language Center. These lessons include simulations in which the student, taking the role of a US soldier, interacts with the local population of a foreign culture in authentic situations, using both English and the local dialect. Digital virtual humans are used to represent not only the soldier, and the role-playing actors; but also a mentor who reviews language and cultural learning points, and provides remediation. Using these prototype lessons as examples, we review the potential advantages of the technology in other training applications.     

An unified VoIP model for workload generation

This paper presents a new model for VoIP workload generation. The novelty of our proposal consists in modeling the sessions by characterizing both the user behavior (session level) and the packet generation for an active call (intra-session level) with easily measured parameters and low computational complexity. This approach also facilitates systematic study of changes in user behavior and voice codec. The session level was modeled by analysis of call-holding time and time interval between successive calls. The model for call-holding time, characterizing the individual user behavior, uses the Pareto type 2 probability distribution. The time interval between calls is obtained from aggregate traffic and can be modeled by exponential probability distribution. Aggregate traffic is obtained by superposition of simultaneous sessions. The data used to characterize the session level were collected at the backbone of two Brazilian telecommunication carriers. The model for intra-session level comprises the characterization of the packet size and the packet inter-arrival time. The intra-session model was based on data generated in a laboratory environment, in order to properly characterize the codec influence on packet generation and to avoid the effects of delay, jitter and loss commonly present in an operational network. Models for constant bit rate and variable bit rate codecs were considered. A simulator was implemented and the results indicate that our model properly mimics the characteristics observed in real traffic and can be used for VoIP modeling and workload generation. Additionally, an application to automate the performance analysis was developed.     

A unified constructive network model for problem-solving

We develop a neural network model that relieves time-consuming trial-and-error computer experiments usually performed in problem-solving with networks where problems, including the traveling salesman problem, pattern matching and pattern classification/learning, are formulated as optimization problems with constraint. First, we specify and uniquely distinguish the model as a set of constituent functions that should comply with restrictive conditions. Next, we demonstrate that it is unified, i.e., it yields most current networks. Finally, we verify that it is constructive, that is, we show a standard method that systematically constructs from a given optimization problem a particular network in that model to solve it.     

A unified model for probabilistic principal surfaces

Principal curves and surfaces are nonlinear generalizations of principal components and subspaces, respectively. They can provide insightful summary of high-dimensional data not typically attainable by classical linear methods. Solutions to several problems, such as proof of existence and convergence, faced by the original principal curve formulation have been proposed in the past few years. Nevertheless, these solutions are not generally extensible to principal surfaces, the mere computation of which presents a formidable obstacle. Consequently, relatively few studies of principal surfaces are available. We previously (2000) proposed the probabilistic principal surface (PPS) to address a number of issues associated with current principal surface algorithms. PPS uses a manifold oriented covariance noise model, based on the generative topographical mapping (GTM), which can be viewed as a parametric formulation of Kohonen's self-organizing map. Building on the PPS, we introduce a unified covariance model that implements PPS (0<α<1), GTM (α=1), and the manifold-aligned GTM (α>1) by varying the clamping parameter α. Then, we comprehensively evaluate the empirical performance of PPS, GTM, and the manifold-aligned GTM on three popular benchmark data sets. It is shown in two different comparisons that the PPS outperforms the GTM under identical parameter settings. Convergence of the PPS is found to be identical to that of the GTM and the computational overhead incurred by the PPS decreases to 40 percent or less for more complex manifolds. These results show that the generalized PPS provides a flexible and effective way of obtaining principal surfaces     

Generating depth of-field effects in virtual reality applications

Computer-generated images without depth-of-field effects always look surreal. A new technique and algorithm are presented which solve the problem of simulating these effects in VR and other real-time applications. The adaptive convolution approach can generate depth-of-field effects in real time using commercially available hardware filters. The picture-processing algorithm proposed can be easily implemented in a visualization pipeline. The final effect can be obtained by iterative filtering with a single hardware convolution filter or with identical 3×3 filters placed as pipeline stages. The speed of this approach makes it suitable for generating images in real-time applications, particularly virtual reality systems. As a postprocessing technique, it also has the advantage of being easily implemented as a final stage of the visualization pipeline in an existing rendering system     

Social facilitation effects of virtual humans

OBJECTIVE: To investigate whether virtual humans produce social facilitation effects. BACKGROUND: When people do an easy task and another person is nearby, they tend to do that task better than when they are alone. Conversely, when people do a hard task and another person is nearby, they tend to do that task less well than when they are alone. This phenomenon is referred to in the social psychology literature as social facilitation. The present study investigated whether virtual humans can evoke a social facilitation response. METHOD: Participants were given different tasks to do that varied in difficulty. The tasks involved anagrams, mazes, and modular arithmetic. They did the tasks alone, in the company of another person, or in the company of a virtual human on a computer screen. RESULTS: For easy tasks, performance in the virtual human condition was better than in the alone condition, and for difficult tasks, performance in the virtual human condition was worse than in the alone condition. CONCLUSION: As with a human, virtual humans can produce social facilitation. APPLICATION: The results suggest that designers of virtual humans should be mindful about the social nature of virtual humans; a design decision as to when and how to present a virtual human should be a deliberate and informed decision. An ever-present virtual human might make learning and performance difficult for challenging tasks.     

Real-time animation of realistic virtual humans

The authors have been working on simulating virtual humans for several years. Until recently, these constructs could not act in real time. Today, however, many applications need to simulate in real time virtual humans that look realistic. They have invested considerable effort in developing and integrating several modules into a system capable of animating humans in real-time situations. This includes interactive modules for building realistic individuals and a texture-fitting method suitable for all parts of the head and body. Animating the body, including the hands and their deformations, is the key aspect of the system; to their knowledge, no competing system integrates all these functions. They also included facial animation, as demonstrated below with virtual tennis players. They have developed a single system containing all the modules needed for simulating real-time virtual humans in distant virtual environments (VEs). The system lets one rapidly clone any individual and animate the clone in various contexts. People cannot mistake virtual humans for real ones, but they think them recognizable and realistic     

Teaching communication skills with virtual humans

At the University of Florida, our research team has worked on applying virtual humans as partners in interpersonal communication scenarios. Our goal is to teach communication skills using virtual humans (VHs). The simulation of an interaction between people to facilitate teaching, training, and testing of communication skills would be a powerful new VR application. Our approach is to employ natural methods of interaction with the VH. We are striving to create an experience similar to two (and eventually more) people talking.