Food intake is a fundamental daily activity. Its relationship to external factors such as social and environmental contexts has been widely investigated in consumer studies. Virtual reality (VR) technology enables the realistic simulation of real-life conditions and, therefore, holds promise as an ecologically valid method for understanding eating intentions and behavior. This study examines whether contextual exposure to VR or text-based storytelling influences food liking and desire in alignment with the presented environment. A total of 117 participants, either online or in a laboratory setting, viewed and rated eight food items varying in calorie density while exposed to one of three conditions: a VR-simulated winter forest, a text-based storytelling winter forest, or a control group without winter cues. Accounting for individual characteristics such as eating habits and body mass index, the results showed that participants' food perceptions -defined here as evaluative responses toward digitally presented food items-followed a similar pattern across both winter forest contexts, with a stronger desire and liking for high-calorie over low-calorie foods. However, the intensity of their perceptual responses varied significantly between conditions. Participants in the VR forest reported greater familiarity, liking, and desire for all food items, as well as a stronger sense of spatial presence, compared to those in the storytelling condition. While VR elicited heightened and more affective responses, both VR and storytelling seem to be useful interventions for assessing contextual effects on product consumption. These findings provide insights into how different modes of contextual presentation shape food perception and encourage the integration of multiple sensory modalities to further bridge the gap between virtual and real food experiences.
Point cloud streaming is a key component of 5G (and beyond) networks, serving as a foundation for holographic communication. While it enables immersive experiences, it also poses significant challenges to network infrastructure. Uncompressed point cloud streams produce Gbps traffic volumes, while compressed streams often suffer from multi-second latencies. Therefore, the feasibility of real-time applications that require ultra-low motion-to-eye latency, such as Network Music Performance and Remote Driving, remains unclear. In this work, we design and implement a novel point cloud streaming tool, based on the Draco encoder, Intel's RealSense SDK, and OpenGL. We deploy our prototype implementation in a private 5G stand alone network and provide an indepth analysis of the latency and throughput of point cloud streaming. Our results show that deploying volumetric streaming services over 5G networks is still a challenging task. However, a combination of simple cost-reduction strategies can bring it significantly closer to feasibility.
Volumetric capture (VolCap) systems are advanced technologies used to capture three-dimensional representations of people, objects, or environments in motion. They provide a means of creating high-quality content essential for professional immersive performing arts experiences. Nascent VolCap systems have not yet resulted in a single standard with few commercial offerings available. In this paper, performer, technician and director user experience (UX) was evaluated using three state-of-the-art VolCap systems in thirteen live performance capture configurations. The performances were volumetrically captured on site, outside of the controlled studio environment. Semi-structured post-experience interviews were conducted. A modified system usability scale was used to record UX. A thematic analysis of the data revealed six themes, strengthened by sentiment analysis. This included significant themes of operational and logistical learnings that resulted in a set of guidelines for use by professionals embarking on volumetric capture of live performances.
This paper investigates the problem of appearance-based virtualization of physical objects in spatial augmented reality, where projected imagery should simultaneously preserve object recognizability, integrate artistic style, and maintain color harmony with a virtual background. We implemented two baseline methods-Average Color, Texture Overlay- and a transformer-based style transfer approach (StyTr.2) and conducted a user study evaluating them across three perceptual criteria: contour clarity (Q1), painterly touch integration (Q2), and color harmony (Q3). The results show that while Average Color excels at contour preservation, and Texture Overlay is superior for stylistic and chromatic integration, only StyTr.2 achieves a perceptual balance across all criteria. Quantitative analysis confirms that StyTr.2 yields the most stable performance with the lowest overall variance. Notably, it received particularly high ratings for painterly touch in scenarios featuring fine-grained texture, such as pointillist styles, suggesting its effectiveness in reproducing style-specific surface qualities.
Network latency presents a critical challenge for interactive multi-user virtual reality (VR) applications, particularly when transmitting the users' full-body motion data to remote collaborators. Such latency can significantly degrade the quality of interaction, leading to a less immersive and less effective user experience. Pose forecasting, which involves predicting future motion sequences based on observed past motions, offers a potential solution. However, it has not yet been widely adopted for compensating latency for full-body motion transmission. In this work, we extend and evaluate two state-of-the-art neural network architectures for predicting future animation sequences, enabling client-side compensation of network latency. To support our methods, we construct a custom motion capture dataset with VR-specific full-body movements, along with an animation-specific data encoding that integrates both joint position and rotation data. Our experimental results show that predictive latency compensation can reduce the positional error of the reconstructed motion by a factor of up to.2.5 and is robust to noisy network connections containing jitter and packet loss. Despite these promising results, we identify several open challenges and outline directions for future work necessary for successful deployment of predictive compensation in real-world VR systems.
