ABSTRACT This paper introduces a reinforcement learning framework for simulating ego‐motorcycle behavior in mixed traffic for animation‐oriented applications. A local path planner uses ray‐based perception to guide navigation, while a tailored reward structure promotes safe interactions and penalizes hazardous maneuvers. To support action selection, the planner also constructs a risk‐probability map that captures nearby vehicles' speeds and distances. Curriculum learning gradually increases traffic complexity during training, enabling the policy to adapt to a wide range of scenarios. Experimental results show that the trained policy produces collision‐free trajectories, responsive deceleration, and well‐regulated maneuvering. It also occasionally exhibits assertive or borderline‐risky interactions that enhance behavioral naturalness. These findings demonstrate that the framework can generate structured yet varied motorcycle motion. Potential applications include entertainment media, educational visualizations, and driver‐training simulations, enriching animated content for learning and engagement.

Large Language Model (LLM) -based conversational agents have become increasingly popular in recent years due to their novel capacity for natural, human-like dialogue interactions. However, mistrust in LLMs persists due to concerns about privacy, the potential for incorrect responses (often referred to as “hallucinations”), and issues related to social bias. Previous AI research shows that anthropomorphic form positively influences users’ perceptions. However, this aspect remains under-explored in LLM-based conversational agent research. Our research features two anthropomorphic forms: embodied and behavioral. Embodied Anthropomorphic Form (EA) encompasses chatbot, chatbot with Text-to-Speech (TTS) , and Embodied Conversational Agent (ECA) interface designs. Behavioral Anthropomorphic (BA) Form involves LLMs instructed with and without Theory of Mind (ToM) principles. In an empirical evaluation, we explored how interplay between BA form and EA form, and vice-versa, affects users’ perceptions of LLM-based conversational agents on trust, anthropomorphism, presence, usability, and user experience. Our findings provide evidence of such effects, offering novel insight into the influence of both anthropomorphic forms on perceived anthropomorphism, presence, usability, user experience, and their positive impact on user trust in LLM-based conversational agents. However, the combined highest (i.e., ECA with ToM behaviors ) and lowest (i.e., Chatbot without ToM behaviors ) levels of both forms result in lower user trust, suggesting a complex relationship between embodiment and ToM behaviors that warrants further investigation.

Understanding visual attention in users during stressful situations with emotional crowds is crucial for the improvement of emergency response, safety, and efficiency. To explore this, we conducted a study examining how the organizational structure and reaction behavior of a virtual emotional crowd affected subjects' visual attention. Participants completed a task in a virtual reality simulation featuring a fire threat. The experiment employed two between-subjects factors (Crowd Type and Reaction Type) and one within-subjects factor (Simulation Phase). The simulation consisted of a normal phase, where participants interacted with virtual human (VHs) pedestrians and vendors in a virtual marketplace, followed by a stressful phase featuring a fire threat within the same setting. The Crowd Type varied between individuals walking alone (Individual Condition) and small groups of 3-4 members (Group Condition). The Reaction Type determined whether the virtual crowd ignored the threat or moved faster to escape. We analyzed subjects' fixation gaze across conditions, revealing that visual attention behavior was significantly influenced by Crowd Type, Reaction Type, and Simulation Phase.

ABSTRACT This paper introduces a reinforcement learning method to simulate agents crossing roads in unsignalized, mixed‐traffic environments. These agents represent individual pedestrians or small groups. The method ensures that agents adopt safe interactions with nearby dynamic obstacles (bikes, motorcycles, or cars) by considering factors such as conflict zones and post‐encroachment times. Risk assessments based on interaction times encourage agents to avoid hazardous behaviors. Additionally, risk‐informed reward terms incentivize agents to perform safe actions, while collision penalties deter collisions. The method achieved collision‐free crossings and demonstrated normal, conservative, and aggressive pedestrian behaviors in various scenarios. Finally, ablation tests revealed the impact of reward weights, reward terms, and key agent state components. The weights of reward terms can be adjusted to achieve either conservative or aggressive pedestrian crossing behaviors, balancing road crossing efficiency and safety.

Evidence from the behavioral psychology literature suggests that in unexpected stressful situations, crowd behavior can have a direct impact on people’s emotions and behaviors. This paper investigates whether this effect is elicited in immersive virtual reality by a crowd of virtual humans (VHs) in an external stressful situation induced by a fire. We conducted a mixed-design study in an experimental scenario where users were asked to collect 10 items from a virtual market. The between-subjects factors were reaction type and agent type, while the within-subjects factor was the simulation phase, which included the normal and stressful situations. There were two reaction types where the VHs showed either active (agitated) or passive (calm) behavior towards the external threat situation. For the agent type, the VHs exhibited either individualistic or group (3-4 member group) behaviors. In a mixed factorial 2(crowd type)x 2(reaction type) x 2(simulation phase) design, we examined the influences of these factors on participants’ emotional and behavioral responses. The results showed that the participants’ task performance and social interaction were affected by the simulated stressor. Negative emotions were elicited when they encountered active VHs and small groups of VHs. Furthermore, a greater change in their interpersonal closeness was experienced when interacting with small groups of VHs, than when interacting with individual VHs.

Summary In this paper, we propose a deep reinforcement‐based approach to generate an animation of an agent using a broom‐like tool to transport a target object. The tool is attached to the agent. So when the agent moves, the tool moves as well.The challenge is to control the agent to move and use the tool to push the target while avoiding obstacles. We propose a direction sensor to guide the agent's movement direction in environments with static obstacles. Furthermore, different rewards and a curriculum learning are implemented to make the agent efficiently learn skills for manipulating the tool. Experimental results show that the agent can naturally control the tool with different shapes to transport target objects. The result of ablation tests revealed the impacts of the rewards and some state components.

This study investigated the extent that the non-verbal behaviors of virtual humans (VHs) and their socio-demographic attributes altered users' collision avoidance behaviors in Virtual Reality (VR). Users interacted with VHs representing different levels of ethnicities and gender, exhibiting different conditions of physical movement, and gaze behaviors. The VHs were depicted in three major ethnic conditions namely Asian, Caucasian, and Black. The physical movement states of the VHs were either static in the path of the user or walking toward the user in the opposite direction. The non-verbal gaze behavior of the VHs was either direct gaze or averted gaze. We used an HTC Vive tracking system to track users' performing real walking while we collected objective measures (i,e., continuous gaze, fixation gaze, clearance distance, and travel length), and subjective variables (i.e., game experiences and social presence). The results showed that the ethnicity of the VHs significantly impacted the gaze behavior of the users, and the gender of the VHs affected the user avoidance movement and their reciprocal gaze behavior. Our results revealed that users' physical movement, gaze behaviors, and collision avoidance were moderated by the VHs' perceived ethnicity, gender, and gaze behaviors. Understanding the impact of the socio-demographics attributes of VHs and their gaze behavior on users' collision avoidance is critical for applications in which users are navigating through virtual traffic, crowd, and other inter-personal simulations.

This paper presents a within-subject study to investigate the effects of leading and following behaviors on user visual attention behaviors when collaborating with a virtual agent (VA) during performing transportation tasks in immersive virtual environments. The task was to carry a target object from a location to a predefined location. There were two conditions, namely leader VA (LVA) and follower VA (FVA). The leader gave instructions to the follower to perform actions. In the FVA condition, users played the leader role, while they played the follower role in the LVA condition. The users and the VA communicated via spoken language. During the experiment, participants wore a head-mounted display and performed real walking in a room. In each condition, each participant performed 20 trials of object transportation for different types of objects. Our preliminary results revealed significant differences in the user visual attention behaviors between the follower and leader VA conditions during the transportation tasks.

Abstract This article presents a reinforcement learning‐based approach to generate animation in which two agents use a rope to collaboratively transport a block. The challenge is that the agents need to master several skills, including approaching the block, using the rope to wrap around it, and then moving the block to a predefined goal position. We propose several reward terms to learn the transportation policy and the adjustment policy that govern the skills of the agents. Experiment results showed that the proposed approach was able to generate various animations in different settings, including rope lengths, block sizes, and block shapes. An ablation test revealed the effects of the reward terms. We also investigated factors that affected the performance of the two policies.

We examined the effects on users during collaboration with two types of virtual human (VH) agents in object transportation in an immersive virtual environment or virtual reality (VR). The two types of virtual humans we examined are leader and follower agents. The goal of the users is to interact with the agents using natural language and carry objects from initial locations to destinations. In each trial, the follower agent follows a user’s instructions to perform actions to manipulate the object. The leader agent determines the appropriate actions that the agent and the user should perform. We developed a system which enabled users and virtual agents to carry objects in an intuitive manner. We conducted a within-subjects study to evaluate the user behaviors under two conditions: (LVH) interaction with a leader virtual human and (FVH) interaction with a follower virtual human. We found that the participants in LVH required a higher workload than that in FVH. However, the users’ experiences, game experiences, and users’ impressions between the two conditions were not significantly different.

This investigation compared the impact on the users’ non-verbal behaviors elicited by interacting with a crowd of emotional virtual humans (VHs) in native and non-native language settings. In a between-subject experiment, we collected objective metrics regarding the users’ behaviors during interaction with a crowd of VHs who expressed verbal and non-verbal communicative behaviors in response to users’ speech-based interaction. The study presented four VH crowd conditions based on their emotional disposition, namely Positive, Negative, Neutral, and Mixed, in accordance with which VHs exhibited the appropriate behaviors. Participants were tasked with collecting items in a virtual flea market, and they interacted using natural speech-based dialogue with the VHs. The language conditions were collected in the USA and under two different conditions in Taiwan. The participants in the USA group interacted with the VHs in English (a native language for the USA setting); and two different groups in Taiwan interacted with the VHs in either a foreign (English) or native (Mandarin) language, respectively. Results reveal that in the (TMA, TEN) and (UEN, TEN) conditions, native versus non-native language communication can alter social behaviors (e.g., interaction time) of participants towards virtual interlocutors. Our results also revealed strong effects of cultural differences on participants’ non-verbal social behaviors with VHs in the UEN and TMA conditions.

Abstract In this article, we propose a framework to generate cart‐pulling animation using deep learning in a multiagent environment. Mainly, two workers pull a cart using ropes and interact with crowd agents which exhibit following, wandering, and evasion behaviors. The main idea is to train a policy to learn an individual behavior of the workers and crowd agents. Furthermore, the challenge is that as the ropes are flexible, rewards are designed carefully so that the workers pull the ropes in a collaborative and consistent manner. Hence, the workers can pull the cart while avoiding collision with the crowd agents and surrounding static objects. In the stage of animation generation, we assign the policies deliberately to the workers and the crowd agents so that they interact with each other naturally. We conducted experiments on animal characters and the system could produce animations of characters with diverse behaviors.

This paper investigated to what extent social interactions and empathy of users could be induced when different control mechanisms were used in an asymmetric collaboration. We conducted a user study to explore the user experience under one decentralized and two centralized control conditions via using the proposed two-player asymmetric collaborative bodily play, LAGH, which supports perspective-taking through the integration with the first- and second-perspectives and shared objects. The two players have complementary views and controls to each other in an immersive environment. The results indicate that participant pairs were encouraged by the asymmetric collaboration interface to share their emotional and physiological perspectives with each other. When their control abilities were balanced, they were more motivated to perform information sharing and interact with each other, thereby enhancing closeness and stimulating empathy. Furthermore, users could improve the collaboration efficiency.

In this contribution, we empirically investigated the effect of small talk on the users' non-verbal behaviors and emotions when users interacted with a crowd of virtual humans (VHs) with positive behavioral dispositions. Users were tasked with collecting items in a virtual marketplace via natural speech-based dialogue with a crowd of virtual pedestrians and vendors. The users were able to engage in natural speech-based conversation in a predefined corpus of small talk content that covered various commonplace small talk topics such as conversations about the weather, general concerns, and entertainment based on similar real-life situations. For instance, the VHs with the small talk ability would ask the users some simple questions to make small talk or remind the users of their belongings. We conducted a between-subjects empirical evaluation to investigate whether the user behaviors and emotions were different between a small talk condition and a non-small talk condition, and examined gender effects of the participants. We collected objective and subjective measures of the users to analyze users' emotions and social interaction behaviors, when in conversation with VHs that either possessed small-talk capability or not, besides task or goal oriented dialogue capabilities. Our result revealed that the VHs with small talk capability could alter the emotions and non-verbal behaviors of the users. Furthermore, the non-verbal behaviors between female and male participants differed greatly in the presence or absence of small talk.

This investigation compared the emotional impact caused by a crowd of affective virtual humans (VHs) that communicated in the users' native or foreign language. We evaluated the users' affective reactions to a crowd of VHs that exhibited distinct emotional expressions. A total of four emotions were presented, which were Positive, Negative, Neutral, and Mixed. The VHs performed verbal and non-behaviors accordingly. Under the Mixed condition, the VHs were divided into three groups equally and each group was uniquely assigned one of the three emotions (i.e., positive, negative, and neutral). Users collected ten items from a virtual reality market. To complete the tasks, they interacted using natural speech with the emotional VHs. Three language conditions were investigated: one condition in USA and another two conditions in Taiwan. The group of participants in USA interacted with the VHs in English; and the two groups of participants in Taiwan interacted with the VHs using a foreign (English) language and a native (Mandarin) language respectively. We discovered that the medium of communication or language familiarity had a strong influence on participants' emotional reactions. When participants interacted in a foreign language with VHs with a positive emotional disposition, we found their positive emotional reactions were subdued and negative reactions were elevated. However, this was not the case when participants interacted with VHs in their native language, as their emotional reactions were contingent on the emotional disposition of the VHs.

Abstract This paper proposes a framework that integrates reinforcement learning and blend‐trees to generate animation of multiple agents for object transportation. The main idea is that in the learning stage, policies are learned to control agents to perform specific skills, including navigation, pushing, and orientation adjustment. The policies determine the blending parameters of the blend‐trees to achieve locomotion control of the agents. In the simulation stage, the policies are combined to control the agents to navigate, push objects, and adjust orientation of the objects. We demonstrated several examples to show that the framework is capable of generating animation of multiple agents in different scenarios.

In this contribution we examined the effects on users during interaction with a virtual human crowd in an immersive virtual reality environment. We developed an agent-based crowd model with rich properties including eye gaze, facial expression, body motion, and verbal and non-verbal behaviors. The scenario was a virtual market in which the users needed to gather specific items. In a betweensubjects design, users interacted with a virtual human crowd that showed opposite valenced emotional expressions. There are four conditions in the between-subjects design. These includes different affective virtual crowds namely, Positive, Negative, Neutral, and a Mix condition. The Mix group is defined by a combination of Positive, Negative and Neutral emotional expressive characters. Depending on the specific condition, the virtual humans showed specific verbal and non-verbal behaviors. During the experiment we collected objective measures such as skin electrodermal activity, total time in the simulation, the number of interactions with the agents and performance measures. The subjective measures were the differential emotional survey (DES), and a user experience survey. We reported our findings with an in-depth analysis.

Abstract In this article, we develop a system with a natural language interface to generate animation of small groups in environments with ambient crowds. Our system takes simple sentences of English as input and these sentences are parsed to obtain information about character attributes, behaviors, and locations for constructing situation nodes. These situation nodes form an animation graph for producing an animation of small groups. The interface assists in interactively managing and editing the animation graph. We demonstrate the effectiveness of the system in some examples. The user study results indicate that the proposed system is user‐friendly and flexible in producing animation of small groups in various scenes.

Abstract We develop an interface that bridges crowd simulation and natural language processing techniques so that casual users can produce crowd animation by text input. The interface adopts a parser and a tagger to analyze simple English sentences to convert them into intermediate data structures that encapsulate the essential elements of crowds. There are five stages: preprocessing, parsing input sentences, crowd generation, animation adjustment, and crowd animation. Our system supports basic behaviors including standing, walking, running, escaping, being attracted, and queuing. We conducted a user experience study to evaluate the interface. The results show that the interface is user‐friendly for casual users to produce crowd animation.

Abstract In this article, we present a framework which transfers video to dynamic relief‐structures with water interaction. Our method enables the interaction between video content and waterflows. The key idea is that the dynamic 2.5D structures of the objects in video are constructed and then waterflows are simulated to interact with the 2.5D structures. We performed experiments for still images, and video in which objects had various motion styles. User interaction with our system was also demonstrated. The result of a user study indicated that our method could produce visually acceptable results.

In this paper, we employ two approaches to perform cooperative transportation of boxes in virtual environments with pedestrians. Each box is pushed by two agents. The agents can change their formation while they push the boxes for collision avoidance with the pedestrians. The two approaches are sequence-based and sensor-based. We perform experiments to evaluate the performances of strategies based on these two approaches under different conditions, including the shortest path and the shortest travel time.

We present a system to help designers create icons that are widely used in banners, signboards, billboards, homepages, and mobile apps. Designers are tasked with drawing contours, whereas our system colorizes contours in different styles. This goal is achieved by training a dual conditional generative adversarial network (GAN) on our collected icon dataset. One condition requires the generated image and the drawn contour to possess a similar contour, while the other anticipates the image and the referenced icon to be similar in color style. Accordingly, the generator takes a contour image and a man-made icon image to colorize the contour, and then the discriminators determine whether the result fulfills the two conditions. The trained network is able to colorize icons demanded by designers and greatly reduces their workload. For the evaluation, we compared our dual conditional GAN to several state-of-the-art techniques. Experiment results demonstrate that our network is over the previous networks. Finally, we will provide the source code, icon dataset, and trained network for public use.

Abstract This paper studies the user collaboration experience with proactive and reactive agents in transporting boxes in virtual environments. Two main characters, the avatar and the agent, are controlled by a user and a controller, respectively. The user and the agent communicate with each other by voice. The agent can be proactive or reactive. The user follows the instruction issued by the proactive agent, whereas the user instructs the reactive agent to perform actions. The goal is to transport boxes to goal positions with orientation constraints. We conducted a user study to analyze the behaviors of participants in several aspects, including task completion time, path length, control experience, and co‐presence experience. We report our findings and make suggestions for future development.

Abstract This paper studies users' learning perception in a virtual titration experiment with differentiated instruction. We develop a virtual reality chemistry lab and use leap motion to detect users' hand gestures for operations. Users wear a head‐mounted display and use their bare hands to interact with virtual objects to perform a titration experiment. Our system implements a complete titration process and provides assistance tools for learning and operating virtual items. We report the essential ideas for building the system. We applied differentiated instruction to study the learning effectiveness of users under different learning intensities. Two groups of students with and without chemistry background participated in a user study. Our results indicated that the virtual reality chemistry lab could enhance and promote users' learning confidence under suitable learning intensities.

This paper presents a novel agent-based approach to simulate cooperative tasks in crowd simulation. Our framework consists of two kinds of agents: pedestrians and workers. A cooperative task requires two or more workers to simultaneously perform actions on objects. The workers should cooperate with each other to finish tasks. This framework considers several basic cooperative behaviors including pushing obstacles away from blocked exits, lifting cargos and carrying them to a destination, and pulling objects using ropes. A complex task can be decomposed into several simple tasks and the simple tasks are finished either in a sequential or parallel order. Our framework also simulates crowd behaviors such as workers joining and leaving cooperative tasks, moving objects around obstacles, collision avoidance for agents, and way giving. We demonstrate examples for simulating different cooperative behaviors. Our framework can be applied to movies, crowd evacuation, narrative systems, task planning simulation, and games.

We present an interactive system using one RGB-D sensor, which allows a user to use bare hands to perform sand drawing. Our system supports the common sand drawing functions, such as sand erosion, sand spilling, and sand leaking. To use hands to manipulate the virtual sand, we design four key hand gestures. The idea is that the gesture of one hand controls the drawing actions. The motion and gesture of the other hand control the drawing positions. There are three major steps. First, our system adopts a vision-based bare-hand detection method which computes the hand position and recognizes the hand gestures. Second, the drawing positions and the drawing actions are sent to a sand drawing subsystem. Finally, the subsystem performs the sand drawing actions. Experimental results show that our system enables users to draw a rich variety of sand pictures.

Abstract This paper focuses on simulating agents transporting passive objects in a virtual environment with pedestrians. The agents adopt thepushingmanipulation patterns to transport the passive objects. We propose techniques to achieve two kinds of cooperation. First, the agents transport the passive objects to the goal positions. During the transportation process, the agents may adjust their formation to transport the passive objects effectively. Second, the agents avoid collision with each other and passive objects in an unfreezing manner. Experiment results show that our technique avoids the freezing problem while the agents are transporting multiple objects.

Abstract This paper studies the effects of a virtual crowd with emotion on task completion in a virtual market. The users have to interact with the virtual agents so that the virtual agents are willing to lead the users to specific vendors. Then, the users can buy the required items. The users communicate with the virtual agents via a conversation dialog. The emotional agents' moods are affected by their own interaction (e.g., collision) while they are walking. Furthermore, their moods are also affected by the conversational responses of the users. We adopt a simple 2D mood model and use icons to represent the mood states of the agents. The mood model supports four mood states, which are “Excitement”, “Anger”, “Sadness”, and “Neutral”. We conducted a user study to evaluate the effects of the emotional virtual agents in different aspects, including task completion, time spent in conversation, following distance, and realism of the emotional virtual crowd.

Abstract Navigation fields are a popular data structure for directing the movement of virtual crowds. However, the congestion problem may occur at the corners of objects in using a navigation field because the agents move quite closely to each other. The agents may congest at the corners even though there is room for the agents to move away from the congested regions. In this paper, we propose to place crowd monitors at the corners to collect the data such as the movement direction of crowds and crowd densities. Then the navigation field is adjusted dynamically so that the crowds are led to move away from the congested regions. We also propose a simple data structure for speeding up the collision detection process between agents and objects. Experimental results show that our approach successfully alleviates the congestion problem at the corners.

In this paper, we present a procedural 2.5D water flow simulation technique to animate pictures. For a given picture, we convert it into a 2.5D structure and then perform water simulation for the 2.5D structure (like a shallow-relief sculpture). The proposed approach is divided into three phases: image processing, water simulation and rendering. First, we extract the objects in the picture and compute their 2.5D structure based on image processing techniques. Second, a procedural based approach is adopted to simulate water flows over the 2.5D structure. Finally, we apply a simple ray tracer to render final images. Experimental results show that the simulated water flows naturally over the 2.5D structure of the pictures at real time rates. We also printed out the 2.5D structures of the virtual objects as 3D models for visual comparison.

An evacuation plan helps people move away from an area or a building. To assist rapid evacuation, we present an algorithm to compute the optimal route for each local region. The idea is to reduce congestion and maximize the number of evacuees arriving at exits in each time span. Our system considers crowd distribution, exit locations, and corridor widths when determining optimal routes. It also simulates crowd movements during route optimization. As a basis, we expect that neighboring crowds who take different evacuation routes should arrive at respective exits at nearly the same time. If this is not the case, our system updates the routes of the slower crowds. As crowd simulation is non-linear, the optimal route is computed in an iterative manner. The system repeats until an optimal state is achieved. In addition to directly computing optimal routes for a situation, our system allows the structure of the situation to be decomposed, and determines the routes in a hierarchical manner. This strategy not only reduces the computational cost but also enables crowds in different regions to evacuate with different priorities. Experimental results, with visualizations, demonstrate the feasibility of our evacuation route optimization method.

We present a simple agent-based approach for crowd simulation in a two-dimensional space. Our approach computes a set of collision-free feasible movement directions for each agent. The movement directions of the agent are determined according to a radial view of the agent in the local region. Subsequently, the direction with the least effort is selected for moving the agent. To efficiently compute feasible movement directions for each agent, we discretize the movement direction field of the agent and use a bit string to estimate the set of movement directions. Each bit of the bit string represents a span of movement directions. The proposed approach enables the agents to move around simple objects without using expensive global search techniques. Experimental results show that our approach can simulate agents that can alleviate the collision problem. The agents can move to their destinations efficiently.

Abstract Climbing plants with tendrils show search and coiling behaviour. A tendril searches for a host object and then twines around it. Subsequently, the tendril coils to pull the main stem of the climbing plant close to the host object. Furthermore, the stems may also twine around the host object. In this paper, we propose a procedural approach to incrementally constructing virtual climbing plants with tendrils that mimic such behaviour. We developed several simple rules to guide the construction process. Although our approach is not based on a physical or biological concept, it is fast and efficient in generating climbing plants with tendrils, with acceptable quality. We propose techniques that are useful for enhancing the realism of climbing plants in close‐up view.

We present an agent-based approach to animate microscopic mixed traffic involving cars and motorcycles in complex scenarios, including signalized and nonsignalized road intersections, and traffic jams due to blockage. Based on our new car-following and lateral movement models, our method can reproduce lane-based and nonlane-based traffic behaviors that are commonly seen in urban scenes. Our dynamic routing and intersection procedures enable a user to interactively control the movement of a car, and our system generates the microscopic behaviors of the influenced vehicles accordingly. Experimental results show that our approach can animate appealing microscopic mixed traffic with various behaviors. Our approach will benefit applications in virtual cities, computer games, and driving simulators.

In this paper, we develop a physics based approach which enables users to use a brush for manipulating water. The users can drip and drag water on a non-absorbent hydrophilic surface to create water artworks. We consider factors, such as cohesive force and adhesive force, to compute the motion of water. Water drops with different shapes can be formed. We also develop a converter for converting input pictures into water-styled pictures. Our system can be applied in advertisements, movies, games, and education.

In this paper, we developed a system which enables users to draw pictures of sand and water on a two-dimensional surface in a virtual environment. By mixing water and sand for artwork creation, we can create different styles of pictures. Our system has three major subsystems: water animation system, sand animation system and rendering system. In the simulation part, we compute the motion of sand and water based on the interactions for water-water, sand-water and sand-sand. In the rendering part, we attempt to balance the rendering quality and the computation cost. We render water using a ray tracing method and the sand is rendered with shadow. Thus, the rendered results of water and sand look like 2.5D. We tested our system and the experimental results showed that the users could create a variety of pictures of sand and water.

Abstract In this study, we develop a system for interactive snow simulation and rendering. Snow is modeled as a hybrid structure that handles movable snow and static snow separately. We develop a simple approach to convert between these two types of snow. Movable snow is represented by a set of particles, whereas static snow is modeled as grid cells. For a piece of movable snow (e.g., snowflake), the particles are connected using springs. Thus, we can model snow as a type of brittle material, such as a snow pile split into chunks of smaller snow pieces after a collision. To render the snow, we adopt a shell structure that has a series of concentric, semitransparent, textured shells. We applied our system to several examples, with the rendered snow appearing similar to real snow. Copyright © 2015 John Wiley & Sons, Ltd.

Abstract In this paper, we propose a method for using particle swarm optimization (PSO) to compute optimal guidance paths for various crowd densities in an agent‐based crowd simulation. The inputs of our system are guidance paths that provide hints for the movement directions of agents. Input guidance paths may not be located correctly (e.g., leading to congestion or high traveling cost); therefore, our method adjusts the guidance paths by using PSO. We consider several factors for evaluating the quality of a guidance path, including the average traveling time and interaction distance between agents. We apply our method in several examples. Experimental results show that our method can compute adaptive guidance paths for various crowd densities. Our system can simulate organized crowds that move in directions specified by the guidance paths. Copyright © 2015 John Wiley & Sons, Ltd.

The radial view-based culling (RVBC) method has been presented for continuous self-collision detection to efficiently cull away non-colliding regions. While this technique mainly relies on the segmented clusters of the reference pose and the associated fixed observer points, it has several drawbacks during the animation and the reduced cost of executing collision detection is limited. We thus present a modified framework to improve the culling efficiency of RVBC. At the preprocessing stage, we segment the closed deformable mesh according to not only the attached skeleton but also the triangle orientations, in order to minimize the collision checks of triangles in a cluster. At the runtime stage, we dynamically merge adjacent clusters and update the positions of observer points if the merged shape is nearly convex. This strategy minimizes the number of triangles in different clusters that required collision check. Our framework can be easily integrated with bounding volume hierarchies to boost the culling efficiency. Experimental results show that our framework achieves up to 5.2 times speedup over the original RVBC method and even more times over the recent techniques.

In this paper, we present an interactive sand art drawing system using Kinect. Our system adopts a vision-based bare hand detection method which effectively detects the hand position and recognizes the hand gestures. Then, the corresponding sand manipulation actions are performed. Our system supports the common sand drawing functions, such as sand erosion, sand sand spilling, and sand leaking. To use hands to draw virtual sand on a drawing plane, we design four key gestures which enable drawing manipulation actions for sand. The basic idea is that the gesture of one hand controls the drawing action. The motion and gesture of the other hand controls the drawing position. Our preliminary experimental results show that our system enables sand drawing with bare hands using Kinect. A user study indicates our system is useful to sand art drawing.

Abstract In this paper, we propose a new continuous self‐collision detection (CSCD) method for a deformable surface that interacts with a simple solid model. The method is developed based on the radial‐view‐based culling method. Our method is suitable for the deformable surface that has large contact region with the solid model. The deformable surface may consist of small round‐shaped holes. At the pre‐processing stage, the holes of the deformable surface are filled with ghost triangles so as to make the mesh of the deformable surface watertight. An observer primitive (i.e. a point or a line segment) is computed so that it lies inside the solid model. At the runtime stage, the orientations of triangles with respect to the observer primitive are evaluated. The collision status of the deformable surface is then determined. We evaluated our method for several animations including virtual garments. Experimental results show that our method improves the process of CSCD.

We propose a new method for auto stylized sand art drawing. Our system is built on top of our sand drawing system which is mixed with a particle system and a height map. An image is input and the patterns of the image is analyzed. The contours of the image are computed and they are converted into strokes. In sand art drawing, artists can use multiple fingers to draw the strokes which are symmetric or parallel at the same time. Our system detects such strokes and groups them according to geometric rules. After analyzing the image and collecting a set of strokes, we enter the drawing stage. In the drawing stage, our system automatically performs sand spilling, drawing strokes, leaking, erosion, pinch spilling and pinch erosion.

Water drops and water flows exhibit interesting motion behaviors and amazing patterns on the surfaces of objects, such as leaves of plants and glass panes. Water drops and water flows are commonly seen in a rainy day. A water drop contains a small amount of water. The motion of a water drop is affected by various factors, including gravity, surface tension, cohesion force and adhesion [Zhang et al. 2012]. The situation is more complicated when we consider the roughness of the surface, surface purities and etc. Kaneda et al. [1993] proposed a discrete model of a glass plate for simulating the streams from the water droplets. The glass plate is divided into a grid. A water droplet is represented as a particle. The law of conservation of momentum is applied for merging droplets. A simple ray tracing technique is adopted for rendering the water droplets that are represented as spheres.

We present a novel radial-view-based culling method for continuous self-collision detection (CSCD) of skeletal models. Our method targets closed triangular meshes used to represent the surface of a model. It can be easily integrated with bounding volume hierarchies (BVHs) and used as the first stage for culling non-colliding triangle pairs. A mesh is decomposed into clusters with respect to a set of observer primitives (i.e., observer points and line segments) on the skeleton of the mesh so that each cluster is associated with an observer primitive. One BVH is then built for each cluster. At the runtime stage, a radial view test is performed from the observer primitive of each cluster to check its collision state. Every pair of clusters is also checked for collisions. We evaluated our method on various models and compared its performance with prior methods. Experimental results show that our method reduces the number of the bounding volume overlapping tests and the number of potentially colliding triangle pairs, thereby improving the overall process of CSCD.