Redirected walking (RDW) enables users to walk naturally in a virtual environment (VE) that is larger than the physical environment (PE). However, relying entirely on redirected walking to navigate large VEs can be challenging. Teleportation, on the other hand, is a popular locomotion technique in VR as it enables rapid travel over long distances without inducing VR sickness. It is thus a practical approach to combine RDW and teleportation for VR navigation, using RDW for nearby destinations and teleportation for distant targets. This study introduces Teleportation with Reorientation (TRO), a technique that inlays reorientation into teleportation to enhance the performance of redirected walking. TRO applies an orientation offset immediately after teleportation, which users counteract by physically turning in place, thereby creating an opportunity to realign their heading to a favorable direction in the PE. A simulation-based evaluation and a user study were conducted to compare TRO against regular teleportations, all used in conjunction with RDW. The results demonstrated that TRO significantly improved the performance of redirected walking compared to regular teleportation. Regarding the overall navigation experience, even though TRO resulted in a higher workload and lower usability because it uses a perceptible reorientation strategy, most participants still considered TRO to be acceptable. In addition, participants indicated that their likeability and preference ratings for the TRO variant featuring one-step turning were the same as those for regular teleportation. These findings suggest that TRO with one-step turning can function as a viable alternative to regular teleportation for VR navigation when both RDW and teleportation are available.

Redirected walking allows users to naturally walk in virtual environments that are larger than their physical surroundings by subtly steering them around real-world obstacles. Alignment-based controllers, such as Vis.-Poly. and Alignment-Optimized, have demonstrated potential in minimizing collision occurrences. However, the assessment of shape similarity using polygonal area has weakened the performance of these controllers since area similarity fails to differentiate between regions with same area but different shapes. This paper presents a weighted area method for approximating the area of a walkable region in front of the user. The area of a walkable region can be estimated by accumulating the subdivided unit areas that intersect the region. These unit areas are assigned weights based on the user's location and orientation. Moreover, Artificial potential fields (APF) value at the center of each unit can be easily incorporated into the weight computation. In a simulation-based evaluation, the Alignment-Optimized controller, implemented using the weighted area method, both without and with the APF value-dubbed as Alignment-Optimized-WA-(AD) and Alignment-Optimized-WA-(AD+APF), respectively-is compared against top-notch controllers including Alignment-Optimized, Vis.Poly., ARC, Alignment+APF-Optimized, APF-S2T, and APF-S2G. The results reveal that both Alignment-Optimized-WA-(AD) and Alignment-Optimized-WA-(AD+APF) exceed the performance of these cutting-edge controllers.

This empirical evaluation aimed to investigate how size perception differs between OST AR and the real world, focusing on two judgment methods: verbal reports and physical judgments. Using a within-subjects experimental design, participants viewed target objects in different sizes in both AR and real-world conditions and estimated their sizes using verbal and physical judgment methods across multiple trials. The study addressed two key hypotheses: (H1) that size perception in AR would differ from the Real World, potentially due to rendering limitations in OST-HMDs, and (H2) that verbal reports and physical judgments would yield different levels of accuracy due to distinct cognitive and perceptual processes involved in each method. Our findings supported these hypotheses, revealing key differences in size perception between the two judgment methods and viewing conditions. Participants consistently underestimated object sizes when using verbal reports in both AR and real-world conditions, with more pronounced errors in AR. In contrast, physical judgments yielded more accurate size estimates under both viewing conditions. Notably, the accuracy of verbal reports decreased as target sizes increased, a trend that was particularly evident in AR. These results underscore the perceptual challenges associated with verbal size judgments in AR and their potential limitations in applications requiring precise size estimations. By highlighting the differences in accuracy and consistency between verbal and physical judgment methods, this study contributes to a deeper understanding of size perception in OST AR and real-world contexts.

Maps have long been a favored tool for navigation in both physical and virtual environments. As a navigation aid in virtual reality, map content and appearance can differ significantly. In this paper, three mini-maps are addressed: the WiM-3DMap, which provides a standard World-in-Miniature of the city model; the novel UC-3DMap, featuring important landmarks alongside ordinary buildings within the user’s vicinity; and the LM-3DMap, presenting only important landmarks. These mini-maps offer varying levels of building detail, potentially affecting spatial knowledge acquisition performance in diverse ways. A comparative study was conducted to evaluate the effectiveness of WiM-3DMap, UC-3DMap, LM-3DMap, and a baseline condition without a mini-map in spatial tasks such as spatial updating, landmark recall, landmark placement, and route recall. The findings demonstrated that LM-3DMap and UC-3DMap outperform WiM-3DMap in the tasks of spatial updating, landmark placement and route recall. However, the absence of detailed local context around the user may impede the effectiveness of LM-3DMap, as evidenced by UC-3DMap’s superior performance in the landmark placement task. These findings underscore the differences in effectiveness among various mini-maps that present distinct levels of building detail. A key conclusion is that including ordinary building information in the user’s immediate surroundings can significantly enhance the performance of a mini-map relying solely on landmarks.

Many AR applications require users to perceive, estimate and calibrate to the size of objects presented in the scene. Distortions in size perception in AR could potentially influence the effectiveness of skills transferred from the AR to the real world. We investigated the after-effects or carry-over effects of calibration of size perception in AR to the real world (RW), by providing feedback and an opportunity for participants to correct their judgments in AR. In an empirical evaluation, we employed a three-phase experiment design. In the pretest phase, participants made size estimations to target objects concurrently using both verbal reports and physical judgment in RW as a baseline. Then, they estimated the size of targets, and then were provided with feedback and subsequently corrected their judgments in a calibration phase. Followed by which, participants made size estimates to target objects in the real world. Our findings revealed that the carryover effects of calibration successfully transferred from AR to RW in both verbal reports and physical judgment methods.

In this contribution, we propose to enhance two distant object manipulation techniques, BMSR (Bimanual Near-Field Metaphor with Scaled Replica) and the classic Scaled HOMER (Scaled Hand-Centered Object Manipulation Extending Ray Casting), via nearfield scaled replica manipulation and viewing. In the proposed Direct BMSR, context replicas are displayed so that the target replica can be manipulated relative to its context, allowing the user to directly manipulate the target replica in their arm’s reach space. Some additional features were implemented to make Direct BMSR an effective interface for manipulating objects from a distance. We proposed Scaled HOMER+NFSRV, which augments Scaled HOMER with a near-field scaled replica view (NFSRV) of the target object and its context, enabling the user to observe how the target replica is manipulated in relation to its context in their arm’s reach space while manipulating it from a distance. We conducted a between-subjects empirical evaluation of BMSR, Direct BMSR, Scaled HOMER, and Scaled HOMER+NFSRV. Our findings revealed that Direct BMSR and Scaled HOMER+NFSRV significantly outperformed BMSR and Scaled HOMER, respectively, in terms of accuracy. This finding highlights the advantages of adding near-field scaled replica viewing and manipulation with respect to distant object manipulation.

Redirected walking (RDW) enables users to walk naturally within a virtual environment that is larger than the physical environment. Recently, several artificial potential field (APF) and alignment-based redirected controllers have been developed and have been demonstrated to significantly outperform conventional controllers. APF Steer-to-Gradient (APF-S2G) and APF Redirected Walking (APF-RDW) utilize the negative gradient and the total force vector, respectively, which are localized to the user's position. These vectors usually point towards the opposite wall when the user is in corridors, resulting in frequent resets within those regions. This paper introduces the APF Steer-to-Target (APF-S2T), a redirected controller that first finds the target sample point with the lowest score in the user's walkable area in both physical and virtual environments. The score of a sample point is determined by the APF value at the point and the distance from the user's position. The direction from the user's position to the target point is then used as the steering direction for setting RDW gains. We conducted a simulation-based evaluation to compare APF-S2T, APF-S2G, APF-RDW, Visibility Polygon-based alignment (Vis.-Poly.) and Alignment-Optimized controllers in terms of the number of resets and the average distance between resets. The results indicated that APF-S2T significantly outperformed the state-of-the-art controllers.

Virtual Reality (VR) is well-known for its use in interdisciplinary applications and research. The visual representation of these applications could vary depending in their purpose and hardware limitation, and in those situations could require an accurate perception of size for task performance. However, the relationship between size perception and visual realism in VR has not yet been explored. In this contribution, we conducted an empirical evaluation using a between-subject design over four conditions of visual realism, namely Realistic, Local Lighting, Cartoon, and Sketch on size perception of target objects in the same virtual environment. Additionally, we gathered participants' size estimates in the real world via a within-subject session. We measured size perception using concurrent verbal reports and physical judgments. Our result showed that although participants' size perception was accurate in the realistic condition, surprisingly they could still tune into the invariant but meaningful information in the environment to accurately estimate the size of targets in the non-photorealistic conditions as well. We additionally found that size estimates in verbal and physical responses were generally different in real world and VR viewing and were moderated by trial presentation over time and target object widths.

Redirected walking allows users to naturally locomote within virtual environments that are larger than or different in layout from the physically tracked space. In this paper, we proposed novel optimization-driven alignment-based and Artificial Potential Field (APF) redirected walking controllers, as well as an integrated version of the two. The first two controllers employ objective functions of one variable, which is the included angle between the user's heading vector and the target vector originating from the user's physical position. The optimized angle represents the physical cell that is best aligned with the virtual cell or the target vector on which the designated point has the minimum APF value. The derived optimized angle is used to finely set RDW gains. The two objective functions can be optimized simultaneously, leading to an integrated controller that is potentially able to take advantage of the alignment-based controller and APF-based controller. Through extensive simulation-based studies, we found that the proposed alignment-based and integrated controllers significantly outperform the state-of-the-art controllers and the proposed APF based controller in terms of the number of resets. Furthermore, the proposed alignment controller and integrated controller provide a more uniform likelihood distribution across distance between resets, as compared to the other controllers.

The goal of this research is to provide much needed empirical data on how the fidelity of popular hand gesture tracked based pointing metaphors versus commodity controller based input affects the efficiency and speed-accuracy tradeoff in users’ spatial selection in personal space interactions in VR. We conduct two experiments in which participants select spherical targets arranged in a circle in personal space, or near-field within their maximum arms reach distance, in VR. Both experiments required participants to select the targets with either a VR controller or with their dominant hand’s index finger, which was tracked with one of two popular contemporary tracking methods. In the first experiment, the targets are arranged in a flat circle in accordance with the ISO 9241-9 Fitts’ law standard, and the simulation selected random combinations of 3 target amplitudes and 3 target widths. Targets were placed centered around the users’ eye level, and the arrangement was placed at either 60%, 75%, or 90% depth plane of the users’ maximum arm’s reach. In experiment 2, the targets varied in depth randomly from one depth plane to another within the same configuration of 13 targets within a trial set, which resembled button selection task in hierarchical menus in differing depth planes in the near-field. The study was conducted using the HTC Vive head-mounted display, and used either a VR controller (HTC Vive), low-fidelity virtual pointing (Leap Motion), or a high-fidelity virtual pointing (tracked VR glove) conditions. Our results revealed that low-fidelity pointing performed worse than both high-fidelity pointing and the VR controller. Overall, target selection performance was found to be worse in depth planes closer to the maximum arms reach, as compared to middle and nearer distances.

We detail the results of two experiments that empirically evaluate the efficiency of spatial selection performance of popular pointing and controller based direct 3D interaction within the participants’ maximum arms reach or personal space in an immersive virtual environment. In the first experiment, targets were presented in a circular configuration of 13 targets in accordance with the ISO 9241-9 Fitts’ Law standard, with a target distance of 5, 9 or 18 cm and target width of 0.65, 0.85, or 1.05cm, and with each circular configuration presented in a random distance of 60%, 75% or 90% of maximum arms length of the participant. The targets were centered at eye level, and in a within-subjects protocol participants selected the targets using a pointer at the tip of the controller or at the tip of their index finger on the dominant hand in the pointing based condition. The second experiment presented each target in a similar width and distance configuration as the previous experiment, but each target also varied in depth from one trial to the next between 60%, 75% or 90% of maximum arms reach. The selections were performed in an HTC Vive head mounted display VR world, using the popular Leap Motion gesture tracking device or the Vive controller. The results of both our experiment taken together revealed that with commodity 3D input devices, participants were more efficient in spatial selection performance in the popular controller based condition, as compared to the pointing based condition. Implications for VR developers and consumers are discussed in the latter sections.

Design discussion is a very important course in architecture education. In this paper, we developed a VR based architecture design discussion system that allows members to visualize and discuss the architectural models and to modify the models during discussion. Since the system is designed to work on top of Rhinoceros and Grasshopper, the object database is updated right after the object modification. Members communicate via voice, object manipulations, and mid-air sketching as well as on-surface sketching in the virtual environment. Several tools have been designed to enhance the sense of presence and to make the discussion more effective. We also developed a rollback mechanism to help users intuitively and quickly revert to a previous state of discussion to make some changes or to start a new direction of discussion. To evaluate the system, we conducted an initial user study with 14 participants to assess the user experience, user impression and effectiveness of the system. The feedback from participants were positive and suggested that the system could be effective and useful for supporting architecture design discussion.

We present a data-driven basketball set play simulation. Given an offensive set play sketch, our method simulates potential scenarios that may occur in the game. The simulation provides coaches and players with insights on how a given set play can be executed. To achieve the goal, we train a conditional adversarial network on NBA movement data to imitate the behaviors of how players move around the court through two major components: a generator that learns to generate natural player movements based on a latent noise and a user sketched set play; and a discriminator that is used to evaluate the realism of the basketball play. To improve the quality of simulation, we minimize 1.) a dribbler loss to prevent the ball from drifting away from the dribbler; 2.) a defender loss to prevent the dribbler from not being defended; 3.) a ball passing loss to ensure the straightness of passing trajectories; and 4) an acceleration loss to minimize unnecessary players' movements. To evaluate our system, we objectively compared real and simulated basketball set plays. Besides, a subjective test was conducted to judge whether a set play was real or generated by our network. On average, the mean correct rates to the binary tests were 56.17 %. Experiment results and the evaluations demonstrated the effectiveness of our system.

In this paper, we present a method to generate realistic defensive plays in a basketball game based on the ball and the offensive team's movements. Our system allows players and coaches to simulate how the opposing team will react to a newly developed offensive strategy for evaluating its effectiveness. To achieve the aim, we train on the NBA dataset a conditional generative adversarial network that learns spatio-temporal interactions between players' movements. The network consists of two components: a generator that takes a latent noise vector and the offensive team's trajectories as input to generate defensive team's trajectories; and a discriminator that evaluates the realistic degree of the generated results. Since a basketball game can be easily identified as fake if the ball handler, who is not defended, does not shoot the ball or cut into the restricted area, we add the wide open penalty to the objective function to assist model training. To evaluate the results, we compared the similarity of the real and the generated defensive plays, in terms of the players' movement speed and acceleration, distance to defend ball handlers and non- ball handlers, and the frequency of wide open occurrences. In addition, we conducted a user study with 59 participants for subjective tests. Experimental results show the high fidelity of the generated defensive plays to real data and demonstrate the feasibility of our algorithm.

Realistic rendering of human faces with makeup is critical for many applications in the 3D facial animation and cosmetic industry. Facial makeup is perhaps one of the most important routines for many females or even for some males. Makeup is a multi-layered process. For example, people usually do the skin care first and smear some cosmetics (such as foundation, blush, lipstick and eye-shadow) on their face. By smearing the cosmetics on the face, facial appearance changes obviously.

Abstract Lighting design plays a crucial role in indoor lighting design, computer cinematograph and many other applications. Computer‐assisted lighting design aims to find a lighting configuration that best approximates the illumination effect specified by designers. In this paper, we present an automatic approach for lighting design, in which discrete and continuous optimization of the lighting configuration, including the number, intensity, and position of lights, are achieved. Our lighting design algorithm consists of two major steps. The first step estimates an initial lighting configuration by light sampling and clustering. The initial light clusters are then recursively merged to form a light hierarchy. The second step optimizes the lighting configuration by alternatively selecting a light cut on the light hierarchy to determine the number of representative lights and optimizing the lighting parameters using the simplex method. To speed up the optimization computation, only illumination at scene vertices that are important to rendering are sampled and taken into account in the optimization. Using the proposed approach, we develop a lighting design system that can compute appropriate lighting configurations to generate the illumination effects iteratively painted and modified by a designer interactively.

ABSTRACT Simulating realistic makeup effects is one of the important research issues in the 3D facial animation and cosmetic industry. Existing approaches based on image processing techniques, such as warping and blending, have been mostly applied to transfer one's makeup to another's. Although these approaches are intuitive and need only makeup images, they have some drawbacks, for example, distorted shapes and fixed viewing and lighting conditions. In this paper, we propose an integrated approach, which combines the Kubelka–Munk model and a screen‐space skin rendering approach, to simulate 3D makeup effects. The Kubelka–Munk model is used to compute total transmittance when light passes through cosmetic layers, whereas the screen‐space translucent rendering approach simulates the subsurface scattering effects inside human skin. The parameters of Kubelka–Munk model are obtained by measuring the optical properties of different cosmetic materials, such as foundations, blushes, and lipsticks. Our results demonstrate that the proposed approach is able to render realistic cosmetic effects on human facial models, and different cosmetic materials and styles can be flexibly applied and simulated in real time. Copyright © 2013 John Wiley & Sons, Ltd.

We present a new framework that combines the hierarchical multi-resolution representation with video-based compression to manage and render large scale time-varying data. In the preprocessing step, the proposed method first constructs a multi-resolution hierarchy using octree structure for each individual time step, and then applies a motion-compensation-based prediction to compress the octree nodes. During rendering stage, the data is decompressed on-the-fly and rendered using hardware texture mapping. The proposed approach eliminates the hierarchical decompression dependency commonly found in the conventional hierarchical wavelet representation methods, which leads to a more efficient reconstruction of data along the time axis. The system provides the user with a spatial region-of-interest (ROI) to adjust the spatial level-of-detail (LOD) selection, and a temporal ROI which is a sub-region only for frequent update during playback. With a suitable control of both ROIs, our system can reach an interactive playback frame rate. This allows the user to observe the dynamic nature of large time-varying data sets.

Abstract We present a novel approach for real‐time rendering of translucent surfaces. The computation of subsurface scattering is performed by first converting the integration over the 3D model surface into an integration over a 2D texture space and then applying importance sampling based on the irradiance stored in the texture. Such a conversion leads to a feasible GPU implementation and makes real‐time frame rate possible. Our implementation shows that plausible images can be rendered in real time for complex translucent models with dynamic light and material properties. For objects with more apparent local effect, our approach generally requires more samples that may downgrade the frame rate. To deal with this case, we decompose the integration into two parts, one for local effect and the other for global effect, which are evaluated by the combination of available methods [DS03, MKB* 03a] and our texture space importance sampling, respectively. Such a hybrid scheme is able to steadily render the translucent effect in real time with a fixed amount of samples.

Abstract Level‐of‐detail modeling is a vital representation for real‐time applications. To support texture mapping progressive meshes (PM), we usually allow the whole PM sequence to share a common texture map. Although such a common texture map can be derived by using appropriate mesh parameterizations that consider the minimization of geometry stretch, texture stretch, or even the texture deviation introduced by edge collapses, we have found that even with a well parameterized texture map, the texture mapped PM still reveals apparent texture distortion due to geometry changes and the nature of linear interpolation used by texture mapping hardware. In this paper, we propose a novel, simple, and efficient approach that adapts texture content for each edge collapse, aiming to eliminate texture distortion. A texture adaptation and its inverse are local and incremental operations that can be fully supported by texture mapping hardware, the render‐to‐texture feature, and the fragment shader. Once the necessary correspondence in the partition of texture space is built during the course of PM construction, the texture adaptation or its inverse can be applied on the fly before rendering the simplified or refined model with texture map. We also propose the mechanism of indexing mapping to reduce blurred artifacts due to under‐sampling that might be introduced by texture adaptation. Keywords: texture mapping progressive meshes, mesh simplification, mesh parameterization, texture distortion Categories and Subject Descriptors (according to ACM CCS): I.3.7 [Computer Graphics]: Three‐Dimensional Graphics and Realism Color, shading, shadowing, and texture

A method is proposed for computing an implicit approximant at a point to a parametric curve or surface. The method works for both polynomially and rationally parameterized curves and surfaces and achieves an order of contact that can be prescribed. In the case of nonsingular curve points, the approximant must be irreducible, but in the surface case additional safeguards are incorporated into the algorithm to ensure irreducibility. The method also yields meaningful results at most singularities. In principle, the method is capable of exact implicitization and has a theoretical relationship with certain resultant-based elimination methods.