With the rapid development of digital media technology, the demand for scientific visualization animation is growing. However, traditional tools such as Blender and Unity face challenges like high barriers to entry and low efficiency. While existing programmatic animation engines improve the efficiency and accuracy of scientific visualization animation production through data-driven approaches, their performance and technical architecture still have room for improvement. Therefore, this project designed and implemented Ranim, a programmatic animation engine based on WebGPU and Rust. The engine utilizes signed distance fields for vector graphics rendering and decouples animation scripts from the core engine, combining the strengths of compiled and interpreted languages to optimize rendering efficiency, cross-platform compatibility, and user experience. Test results show that the engine outperforms existing solutions in most scenarios and across a wide range of platforms.

IntroductionPublic speaking practice in Virtual Reality (VR) provides an immersive and controllable setting, yet many existing systems offer limited adaptive guidance beyond exposure itself. Recent advances in Large Language Models (LLMs) create opportunities for designing coaching interactions that are personalized, context-aware, and socially supportive.MethodsWe designed a VR public speaking system in which embodied virtual judges delivered spoken and textual, performance-contingent coaching based on automatic speech recognition, speech synthesis, gaze-related cues, and structured prompting with ERNIE Bot. We evaluated the system in a two-arm randomized controlled user study with 120 participants, with 60 participants assigned to the adaptive coaching condition and 60 to an attention-matched generic feedback control condition. Participants completed two rounds of VR speeches with repeated pre- and post-speech self-report assessments, followed by matched follow-up sessions at one week and one month.ResultsCompared with attention-matched generic feedback, the adaptive coaching condition was associated with larger reductions in self-reported state anxiety and PRCA-derived public speaking anxiety appraisals across repeated speaking rounds. At follow-up sessions conducted in the same VR environment without additional coaching, lower self-reported anxiety remained observable in the experimental group. Qualitative feedback suggested that participants experienced the agents as supportive, useful, and lower-pressure, while also noting limitations in realism, speech prosody, and interaction richness.DiscussionThe findings provide preliminary evidence that embodied LLM-based coaching can support anxiety-aware VR public speaking practice. However, because the study relied primarily on self-report, did not include physiological outcomes or blinded external performance ratings, and involved repeated exposure to the same VR context, the follow-up findings should be interpreted as maintenance within the VR practice setting rather than evidence of stable trait change, objective performance improvement, or real-world transfer. Future studies should examine whether these self-reported gains generalize to real-world speaking contexts and externally rated public speaking performance.

Distance perception plays a critical role in shaping user experience in Virtual Reality (VR) environments. While auditory distance perception has been extensively studied in virtual acoustics, much less is known about how structured visual reference grids and calibration procedures jointly shape auditory distance perception across different directions in fully immersive VR. In particular, little is known about whether visual reference grids can improve auditory distance judgments across multiple directions and whether calibration with such grids can produce benefits that persist after the grids are removed. To address this issue, we conducted a user study consisting of three sequential phases: pre-test, calibration, and post-test. A multifactorial experimental design was adopted, incorporating 3 (visual reference grid conditions) × 3 (phases) × 5 (sound source distances) × 4 (sound source sectors), with a total of 90 participants. The results revealed that participants achieved more accurate auditory distance judgments, with greater calibration-related improvement, in the left and right sectors than in the front and back sectors. The presence of persistent visual reference grids significantly improved distance judgments across directions. In addition, when visual grids were presented only during the calibration phase, participants showed descriptively improved post-test performance relative to the no-grid baseline in the subsequent visual-reference-free test phase, although this effect was not uniform across all primary mixed-effects outcomes.

Emotion is a complex psychophysiological phenomenon elicited by external stimuli, exerting a profound influence on cognitive processes, decision-making, and social behavior. Emotion recognition holds broad application potential in healthcare, education, and entertainment. With virtual reality (VR) emerging as a powerful tool, it offers an immersive and controllable experimental environment. Prior studies have confirmed the feasibility and advantages of VR for emotion elicitation and recognition, and multimodal fusion has become a key strategy for enhancing recognition accuracy. However, publicly available VR multimodal emotion datasets remain limited in both scale and diversity due to the scarcity of VR content and the complexity of data collection. The shortage hampers further progress. Moreover, existing multimodal approaches still face challenges such as noise interference, large inter-individual variability, and insufficient model generalization. Achieving robust and accurate physiological signal processing and emotion modeling in VR environments thus remains an open challenge. To address the issues, we constructed a VR experimental environment and selected 10 emotion-eliciting video clips guided by the PAD(Pleasure-Arousal-Dominance) model. Thirty-eight participants (N=38) were recruited, from whom electrodermal activity, eye-tracking, and questionnaire data were collected, yielding 366 valid trials. The newly collected dataset substantially extends the publicly available VREED dataset, enriching VR-based multimodal emotion resources. Furthermore, we propose the MMTED model (Multi-Modal Temporal Emotion Detector), which incorporates baseline calibration and multimodal fusion of electrodermal and eye-tracking signals for emotion recognition. Experimental results demonstrate the strong performance of the MMTED model, achieving accuracies of 85.52% on the public VREED dataset, 89.27% on our self-collected dataset, and 85.29% on their combination.

Edge computing has emerged as a promising paradigm for supporting latency-sensitive and computationally intensive applications. In this context, effective server deployment and service provisioning are crucial for optimizing system performance and maximizing platform profit. This paper addresses the joint problem of server deployment and service provisioning in multi-user scenarios, with the goal of minimizing overall cost and user access latency while satisfying constraints such as distance thresholds, resource capacities, and connectivity requirements. To this end, we propose a novel two-stage framework that decouples the problem into two interrelated but relatively independent stages. In the first stage, we aim to minimize both deployment cost and transmission cost by proposing a cost-benefitdriven server deployment strategy. This strategy considers the geographical distribution of users and base stations to determine an appropriate number of servers based on a cost budget and user density. We further design a server deployment algorithm inspired by the K-Median scenario and employ a Coveragebased Local Search (CLS) approach to determine the specific server locations. In the second stage, we focus on optimizing both the cost and user delay through a service provisioning strategy by considering multi-objective optimization. We improve the initialization phase by generating higher-quality initial solutions and propose an intelligent service provisioning strategy oriented toward the Pareto front (PSP), thereby achieving a joint optimization of provisioning cost and service latency. Extensive experimental results demonstrate that the strategy effectively reduces deployment expenses and user response delays, offering improved service quality and enhanced economic benefits for service providers and network operators.

As technology advances, the application of Virtual Reality (VR) in the field of mental health continues to grow, particularly showcasing its unique advantages in the adjustment and treatment of emotional and psychological states. Traditionally, the diagnosis of depression relies on clinical psychologists who assess symptoms through conversation and behavioral observation, supported by standardized psychological assessment tools. These methods rely primarily on subjective reports and the judgment of clinicians, which can lead to inconsistencies and reproducibility problems in the diagnosis results. This study utilizes VR technology combined with physiological measurement tools to explore and analyze the emotional response differences between depressed patients and healthy individuals while watching VR videos with different emotional colors, and attempts to predict the state of depression. Forty-six participants are exposed to three types of emotionally colored videos in a VR environment. The correlation between electrodermal activity (EDA), heart rate, and depression is investigated. The results indicate significant differences in psychological and physiological responses between depressed patients and healthy individuals, suggesting these differences could serve as new biomarkers for the diagnosis and prediction of depression. This study extends the application of VR technology in mental health assessments, providing a scientific basis for more objective diagnostic and predictive methods for depression, potentially enhancing the accuracy and efficiency of diagnosis and prediction.

Cross-city transfer learning aims to apply the knowledge and model from data-rich cities to data-poor cities to solve the cold start problem. Existing methods directly transfer the model constructed from developed cities to underdeveloped cities without considering the similarity between them, which leads to a potential transfer mismatch problem, and in turn, decreases the performance of transfer results. Meanwhile, existing transfer learning methods cannot effectively extract the time series features of the data, resulting in the inability to achieve adaptive positive migration across cities. To solve this problem, we propose a similarity-based cross-city transfer learning method named TransCSM, which embeds the urban similarity into an adaptation transfer learning framework to achieve desired data transfer. Specifically, we first constructed an urban similarity model, which utilizes the urban POI (Point Of Interest) data to group the cities with similar characteristics into the same cluster. Then, we build a feature extractor network, that uses convolution neural network (CNN) and Gated Recurrent Unit (GRU) to extract more representative features of time series data. Afterwards, we build an adaptation transfer learning framework to achieve data transfer within the same city cluster, which ensures the reliability of cross-city data transferring results. Finally, we evaluate our proposed method in many public POI datasets from Baidu Map API, and enormous results have demonstrated that our proposed method can achieve superior performance against state-of-the-art methods.

In recent years, virtual reality (VR) technology has advanced rapidly, with spatial audio playing a key role in enhancing immersion. The Head-Related Transfer Function (HRTF) is central to realistic three-dimensional soundscapes, significantly improving directional perception. However, challenges remain regarding users' ability to accurately recall and comprehend auditory content in virtual environments, impacting overall audio intelligibility, recall, and immersion. This study conducted a 3 (VR scenario) x 3 (sound source type) x 3 (audio mode) factorial experiment with ninety participants to investigate effects on intelligibility, recall, and immersion. Results revealed that participants exhibited superior audio recall for static, moving, and rotating sound sources positioned on the left and right compared to other directions. Additionally, compared to simulated 5.1 and 7.1 surround sound systems, VR spatial audio significantly enhanced user immersion, intelligibility, and recall.

With the popularization and development of virtual reality technology, panoramic video has gradually become one of the mainstream forms of VR technology in various fields. However, the research on the information perception of panoramic video in different media is insufficient. And shortcomings still exist in building information perception and prediction models owing to small samples. This work focuses on users’ perception of multi-source information in panoramic videos with different media. We conducted the experiment (N = 40) to analyze the differences of users’ perception level when viewing panoramic videos using different media (i.e. VR and traditional media). We also studied the correlation between user characteristics and information reception effectiveness. The results show that users’ perception of multi-source information in VR is better than in traditional media, except for sound information. Besides, there is a positive correlation between observational ability, memory, concentration, and spatial perception, whether playing computer games frequently and multi-source information perception.

Virtual reality (VR) interaction safety is a prerequisite for all user activities in the virtual environment. While seeking a deep sense of immersion with little concern about surrounding obstacles, users may have limited ability to perceive the real‐world space, resulting in possible collisions with real‐world objects. Nowadays, recent works and rendering techniques such as the Chaperone can provide safety boundaries to users but confines them in a small static space and lack of immediacy. To solve this problem, we propose a dynamic approach based on user motion prediction named SCARF, which uses Spearman's correlation analysis, rule learning, and few‐shot learning to achieve prediction of user movements in specific VR tasks. Specifically, we study the relationship between user characteristics, human motion, and categories of VR tasks and provides an approach that uses biomechanical analysis to define the interaction space in VR dynamically.We report on a user study with 58 volunteers and establish a three dimensional kinematic dataset from a VR game. The experiments validate that our few‐shot learning model is effective and can improve the performance of motion prediction. Finally, we implement SCARF in VR environment for dynamic safety boundary adjustment.

The development of augmented reality (AR) has brought innovation to the working mode in multiple fields to improve the efficiency of multi-party cooperation and the safety rate of operations. However, when users use AR technology, they will inevitably suffer from cybersickness, which may cause symptoms such as nausea, dizziness, eye fatigue, and headaches, severely reducing their experience. Still, there is limited research on AR cybersickness, compared to cybersickness in virtual reality (VR). The causes and symptoms of cybersickness in AR and VR are different, therefore some conclusions in VR may not be applicable. Also, the current research scenarios related to AR cybersickness lack diversity, which is not conducive to the study of the causes of cybersickness. The paper mainly studies the cybersickness in AR railway inspection scenarios, including the causes (latency and the number of interactions) and the symptoms. 54 volunteers were recruited for the experiment, and we defined three latency types and three interaction types based on related interaction criteria. Then we designed a 3 * 3 experiment with latency and interaction quantity as variables. We conducted statistical analysis on the ARSQ data and experimental data. The results show that the severity of cybersickness in AR railway inspections is significantly positively correlated with the number of interactions and latency. Oculomotor is also significantly correlated with the two variables, while the severity of disorientation is not significantly correlated with latency. Besides, it was also observed that in AR, the severity of oculomotor is much higher than disorientation.

Cybersickness remains a pivotal factor that impacts user experience in Augmented Reality (AR). Research probing into the relationship between AR reading tasks and cybersickness, particularly focusing on text display patterns and user characteristics, has been scant. Moreover, the influence of cybersickness on searching ability and the broader spectrum of user experience has not been rigorously tested. Recent investigations have aimed to pinpoint the variables that contribute to cybersickness during AR reading sessions. In one such study, 40 participants underwent a series of controlled experiments with randomized text display patterns, including variations in text speed and text movement modes. Post-experiment, participants completed a questionnaire that helped quantify their experiences and the degree of cybersickness encountered. The data highlighted that satiety, text speed, and text movement mode are significant contributors to cybersickness. When participants experienced higher levels of cybersickness, font color stood out as a particularly influential factor, whereas gender differences seemed to affect the onset of cybersickness more noticeably at lower levels. This study also drew attention to the impact of cybersickness on search ability within AR environments. It was noted that as cybersickness intensity increased, search ability was markedly compromised. In sum, the research underscores the importance of text display patterns and user characteristics, such as past AR experience, in understanding cybersickness and its detrimental effects on user experience and search ability, particularly under conditions of intense cybersickness.

Virtual reality (VR) utilizes computer vision, artificial intelligence and other techniques to enable interaction between users and virtual environments. In order to solve the problems of human joint localization based on single device, multi-device data fusion technology has been adopted. In this paper, a multi-device data fusion method is proposed based on HTC Vive and Kinect. Firstly, two devices are utilized to separately capture motion data of human joints and the two sets of data are aligned temporally and unified in coordinates. Then the weights are respectively assigned to the two sets of data based on the different location of the human body. Next, particle filtering is adopted to combine the two sets of data. Finally, a bidirectional long short-term memory (Bi-LSTM) neural network model is deployed, where the bone length loss is incorporated into the loss function to further improve the localization accuracy. Experiment results show that the localization accuracy of the proposed multi-device data fusion-based localization method outperforms that of the single device method.

This paper investigates the application of Virtual Reality Exposure Therapy (VRET) to treat agoraphobia, focusing on two pivotal research questions derived from identified gaps in current therapeutic approaches. The first question (RQ1) addresses the development of complex VR environments to enhance therapy's effectiveness by simulating real-world anxiety triggers. The second question (RQ2) examines the differential impact of these VR environments on agoraphobic and nonagoraphobic participants through rigorous comparative analyses using t-tests. Methodologies include advanced data processing techniques for electrodermal activity (EDA) and eye-tracking metrics to assess the anxiety levels induced by these environments. Additionally, qualitative methods such as structured interviews and questionnaires complement these measurements, providing deeper insights into the subjective experiences of participants. Video recordings of sessions using Unity software offer a layer of data, enabling the study to replay and analyze interactions within the VR environment meticulously. The experimental results confirm the efficacy of VR settings in eliciting significant physiological and psychological responses from participants, substantiating the VR scenarios' potential as a therapeutic tool. This study contributes to the broader discourse on the viability and optimization of VR technologies in clinical settings, offering a methodologically sound approach to the practicality and accessibility of exposure therapies for anxiety disorders. This paper investigates the application of Virtual Reality Exposure Therapy (VRET) to treat agoraphobia, focusing on two pivotal research questions derived from identified gaps in current therapeutic approaches. The experimental results confirm the efficacy of VR settings in eliciting significant physiological and psychological responses from participants, substantiating the VR scenarios' potential as a therapeutic tool. This study contributes to the broader discourse on the viability and optimization of VR technologies in clinical settings, offering a methodologically sound approach to the practicality and accessibility of exposure therapies for anxiety disorders. image

Nowadays, VR technology has play an indispensable role in promoting the construction of the Metaverse. In this study, we tried to avoid users' body physical collisions within the safety boundary in VR scenes. Three basic motion event categories are defined according to the common motions of users. We segmented users data into three-dimensional coordinates, extracted the motion range features, and constructed the relational data set. Furthermore, we used statistical methods to study the correlation between user characteristics and motion range categories, in order to explore and analyze the possibility of optimizing the safety motion range. Finally, we deployed our safety boundary systems in the virtual reality environment.

Deep learning techniques continue to be used in various applications in recent years. However, when it is difficult to obtain adequate training samples, the performance of the depth model will degrade. Although few-shot learning and data enhancement techniques can relieve this dilemma, the diversity of real data is too large to simulate. To tackle this challenge, we study a novel method, Data Augmentation Scheme For Few-Shot Object Detection (DA-FSOD), to improve the efficiency of model training on visual tasks. Specifically, to expand data augmentation space, we build a data augmentation operation pool (DAOP) based on several common-applied image process operations. Then we propose a novel data augmentation scheme, the series and parallel connection scheme, which superimposes the effects of different operations to generate diverse variants. To further explore and utilize the deep feature information, we leverage the semantic information of input image in model and propose imposed semantic data augmentation which augments training set semantically via deep features of augmented variants. The proposed method successfully enhanced the model performance. We validated our approach using extensive experiments on the domain of few-shot object detection. The results showed remarkable gains compared to state-of-the-art methods.

In recent years, virtual reality (VR) technology has advanced significantly, and VR spatial audio technology has greatly enhanced the immersive experience for users. Despite the progress made in VR spatial audio technology, the acceptance and perception of sound source by users in VR environments remains relatively low, which has a negative impact on the immersive experience. The field of virtual reality user perception is currently under-researched, and existing studies have not adequately addressed The impact of sound source orientation on user perception and immersion and the categorization of sound source types. Therefore, this paper mainly discusses the following two questions: Do differences in sound source orientation and sound mode affect user perception in virtual reality environments? Is the user’s immersion in a virtual reality environment related to the sound mode? To answer these questions, this study recruited 66 participants to conduct experiments, questionnaires, and semi-structured interviews. Using statistical methods, we analyzed the collected dataset and found that sound source orientation and type significantly influence user perception. The results demonstrate that users’ perception and immersion in VR spatial audio mode are significantly better than those in traditional 5.1 and 7.1 surround sound modes.

Virtual Reality (VR) is a new technology that emerged and rapidly developed in the 20th century. Its scenes have an extremely high degree of realism, providing users with an immersive sensory experience. However, many users may experience cybersickness, which can cause symptoms such as dizziness and nausea, severely affecting users’ VR experience and greatly limiting the development and popularity of VR. Based on previous research results, we find a lack of research on cybersickness related to movement patterns and light pollution scenes. Therefore, this paper mainly discusses the following two questions: Do different movement patterns and light pollution scenes cause users to experience cybersickness? What is the correlation between human factors and the degree of cybersickness in movement patterns and light pollution scenes? In response to the questions, a total of 41 participants were recruited for the research. The results show that compared to straight and straight with sudden brakes, the pattern of straight with turns makes users have a higher degree of cybersickness. In white light scenes, people tend to feel dizzier at a moderate level. While in colored scenes, there exists a linear relationship between light intensity and cybersickness. Moreover, user characteristics also have an impact on cybersickness. It shows that individuals who have experienced motion sickness during driving are more likely to experience dizziness during movement scenes in VR. People who pay more attention to environmental light in daily life tend to feel dizzier in white light scenes in VR and those who consider light pollution in their place of residence to be serious also tend to experience higher levels of cybersickness in light pollution scenes.

Vehicular mobility model plays an important role in high-quality virtual traffic flow simulation. The neural network based method is capable to extract individual mobility models from traffic data, however, scenario simulation based on these models tends to go wild due to the error accumulation caused by the mixture of data from different driving states. In this paper, prior knowledge in vehicle kinetics will be introduced to the data-driven mobility model. A hierarchical framework is proposed with a Hidden Markov Model (HMM) layer to distinguish the driving states and a mobility model trained by specified data, along with an adaptation method to achieve the probabilistic representation of the kinetic car-following model. The partial knowledge on the indicated car-following driving state will also help initialize the HMM. The HMM delivers the temporal dependency and the labeling procedure provides more specified and precise data to the neural network. Finally, several experiments will be given to demonstrate the validity of the proposed method.

Virtual Reality(VR) technology builds a simulation environment through a computer,provides users with a three-dimensional dynamic view,and enhances the user's sensory experience,so that the user will get an immersive sense of immersion.With the rise and continuous development of virtual reality technology,people's visual and auditory experience has made great progress.With the development of multimedia technology,panoramic video emerges gradually.Compared with ordinary video,panoramic video has a wider viewing angle and richer visual information.The wide application of virtual reality technology and the development of panoramic video technology make virtual reality panoramic video(VR Video) become one of the most popular and concerned VR services.The information perception and acceptance behavior of users in the virtual reality environment have also been affected.Based on SMOTE algorithm,Bayesian network,logistic regression,and other statistical analysis methods and Few-shot Learning algorithm,this paper compares the differences in users' information memory degree,sense of reality,sense of participation,and other aspects when using VR head-mounted display and ordinary display iPad to watch panoramic video respectively.To explore the difference between the user's information acceptance effect and the sense of presence in the virtual reality environment and the traditional media environment.Experiments show that users with virtual reality headsets score an average of about 0.786 for message acceptance,compared with 0.634 for iPad users with regular displays.Among them,the score of positive information and non-positive information receiving effect in the virtual reality environment is 1.626 times and 1.245 times that of in the traditional environment respectively.However,it is noteworthy that in video A,which has the longest average scene duration,there is no significant difference in the information acceptance effect.In addition,the length of the user's visual residence time has a positive impact on the acceptance effect of information.After verifying that the sense of presence can be subdivided into the sense of reality and sense of participation,this paper shows that the score of users for the sense of presence in the iPad environment is always lower than that of in a VR environment through feature engineering and classification decision tree.After that,the correlation coefficient of the least square method is used to prove that presence has a positive effect on user information acceptance.Meanwhile,with the help of the Few-shot Learning algorithm,the average information memory number of video A－D is 9.20,9.13,8.83,and 10.57 when the sense of presence is strong,while the average information memory number of video A－D is 8.53,6.80,7.14 and 7.66 when the sense of presence is weak.It is proved that the user's sense of presence during information perception is beneficial to the user to obtain a better information acceptance effect.