1,311 research outputs found
EDM-Research/SLF4Web: v0.1.0 - source code IC3D 2021
No full textSource code for the following scientific publication:
Hendrik Lievens, Maarten Wijnants, Brent Zoomers, Jeroen Put, Nick Michiels, Peter Quax, and Wim Lamotte. Adaptive Streaming and Rendering of Static Light Fields in the Web Browser. In: Proceedings of the 12th International Conference on 3D Immersion (IC3D), 2021.
Additional light field datasets can be downloaded from https://doi.org/10.5281/zenodo.573052
Progressive Network Streaming of Textured Meshes in the Binary glTF 2.0 Format
No full textThe glTF 2.0 graphics format allows for the API-neutral representation of 3D scenes consisting of one or multiple textured meshes. It is currently adopted as one of two file formats for 3D asset interop-erability by the Metaverse Standards Forum. glTF 2.0 has however not been designed to be streamable over the network; instead, glTF 2.0 files typically first need to be downloaded fully before their contents can be rendered locally. This can lead to high start-up delays which in turn can lead to user frustration. This paper therefore contributes a methodology and associated Web-based client, implemented in JavaScript on top of the three.js rendering engine, that allows to stream glTF 2.0 files from a content server to the consuming client up to the level of individual glTF bufferviews. This in turn facilitates the progressive client-side rendering of 3D scenes, meaning that scene rendering can already commence while the glTF file is still being downloaded. The proposed methodology is conceptually compliant with the HTTP Adaptive Streaming (HAS) paradigm that dominates the contemporary market of over-the-top video streaming. Experimental results show that our methodology is most beneficial when network throughput is limited (e.g., 20Mbps). In all, our work represents an important step towards making 3D content faster accessible to consuming (Web) clients, akin to the way platforms like YouTube have brought universal accessibility for video content
Adaptive Web-Based VR Streaming of Multi-LoD 3D Scenes via Author-Provided Relevance Scores
No full textThe growing storage requirements of 3D virtual scenes, combined with the increased heterogeneity of consumption devices, trigger the need for novel, on-demand streaming techniques of textured meshes. This paper proposes a way to perform relevance-aware Adaptive Bit-Rate (ABR) scheduling using MPEG-DASH, tailored to VR consumption in the web browser. Scene authors can annotate the relative importance of scene assets to optimize scheduling decisions. Our approach outperforms the state-of-the-art (measured using the MS-SSIM metric) across different scene complexities and network configurations, and is found to be most beneficial when scene complexity is high and network conditions are relatively poor.IEEE; IEEE Comp Soc; Virbela; Tecnico Lisboa; Immers Learning Res Network; Qualcomm; Vicon; HitLabNZ AIGI; Microsoft; Appen; Facebook Real Labs Res; XR Bootcamp; NSF; Fakespace Lab
Empowering Adaptive Learning in VR Assembly Training Using Real-time Performance Tracking
No full textVirtual Reality (VR) provides unique opportunities for creating immersive, per-sonalized and responsive learning environments through advanced features like hand and eye tracking. However, traditional VR training often lacks the flexibility to accommodate diverse learning styles. Personalization of training is crucial in industrial assembly, where user profiles and levels of expertise vary greatly. This paper introduces a novel solution for adaptive learning in VR focused on assembly knowledge training, using hand and eye tracking to deliver real-time feedback and individually adjusted learning paths. We applied our approach to two realistic assembly cases to evaluate its practical application. We hope to inspire future research further to explore and refine this adaptive approach, contributing to developing more flexible and effective VR-based training solutions for the manufacturing industry.ThisresearchwassupportedbyFlandersMake,thestrategicresearchcentreforthemanufacturingindustry, inthe projectSKILLEDWORKFORCE
Does One-Size Training Fit All? Evaluating Adaptive Learning for VR Assembly Training
No full textVirtual Reality (VR) is gaining popularity and is increasingly adopted across various industries for its potential to deliver immersive and effective skill development. However, we observe that VR training often follows a one-size-fits-all approach. Trainings typically do not adapt to to individual skill levels, which is particularly important in industrial assembly, where user profiles and expertise levels vary widely. To address this, we applied the concept of adaptive learning to VR assembly training, enabling the system to dynamically provide assistance levels when users struggle and gradually reduce support as their proficiency increases. This paper investigates the learning performance and subjective impact of two types of such adaptive approaches and a non-adaptive variant in a VR user study with 36 participants. The results show that adaptive training significantly enhances user experience and reduces perceived workload. At the same time, adaptive VR learning is found to have a positive impact on learning performance (quantified as a reduced number of assembly mistakes after training). In summary, our findings underscore the potential of applying adaptive learning approaches in VR. To guide future research, we propose guidelines to support the practical adoption of adaptive learning in VR training in manufacturing and beyond.This research was supported by Flanders Make, the strategic research centre for the manufacturing industry, in the project SKILLEDWORKFORCE
Split & Dual Screen Comparison of Classic vs Object-based Video
No full textOver-the-top (OTT) streaming services like YouTube and Netflix induce massive amounts of video data, hereby putting substantial pressure on network infrastructure. This paper describes a demonstration of the object-based video (OBV) methodology that allows for the quality-variant MPEG-DASH streaming of respectively the background and foreground object(s) of a video scene. The OBV methodology is inspired by research into human visual attention and foveated compression, in that it allows to adaptively and dynamically assign bitrate to those portions of the visual scene that have the highest utility in terms of perceptual quality. Using a content corpus of interview-like video footage, the described demonstration proves the OBV methodology's potential to downsize video bitrate requirements while incurring at most marginal perceptual impact (i.e., in terms of subjective video quality). Thanks to its standards-compliant Web implementation, the OBV methodology is directly and broadly deployable without requiring capital expenditure.Maarten Wijnants is funded by a VLAIO Innovation Mandate (project number HBC.2016.0625), co-sponsored by Androme. Sven Coppers is funded by the Special Research Fund (BOF) of Hasselt University (R-8150). We thank Davy Vanacken for his methodological advice
Cross-layer metrics sharing for QUICker video streaming
No full textQUIC is marketed to hold many advantages over TCP. However, preliminary experimentation has shown that simply running contemporary HTTP Adaptive Streaming (HAS) implementations over QUIC does not improve but actually hurts streaming performance compared to a traditional TCP deployment. We argue that this behavior can be attributed to the amount of TCP specialization that HAS Adaptive BitRate (ABR) algorithms have received over the years. In contrast to TCP (which can be regarded as a "black box"), QUIC actually encompasses all the necessary tools to empower streaming performance optimization (e.g., definition of custom congestion control algorithms, access to transport-layer metrics in the application layer). This however comes at the expense of added complexity which in turn could lead to misinterpretations of the root causes of suboptimal streaming performance. To facilitate research on HTTP adaptive bitrate streaming over QUIC, in this paper we propose a solution towards jointly visualizing transport-and application-layer metrics to allow for a better understanding of HAS streaming performance over various types of transport protocols (i.e., TCP versus QUIC). We see the work presented in this paper as an important stepping stone towards cross-layer optimization of HAS ABR performance to achieve a better overall Quality of Experience (QoE) for streaming users
Cross that boundary: Investigating the feasibility of cross-layer information sharing for enhancing ABR decision logic over QUIC
No full textWith HTTP Adaptive Streaming (HAS), client-side Adaptive Bi-trate (ABR) algorithms drive the (quality-variant) scheduling and downloading of media segments. These ABR algorithms are implemented in the application layer and can therefore base their logic only on relatively coarse and/or inaccurate application-layer met-rics. The recently standardized QUIC transport protocol has many userspace implementations, which paves the way for cross-layer optimizations by exposing transport-layer metrics to application-layer algorithms. In this paper, we investigate whether the availability of fine-grained transport-level throughput metrics can positively impact the operation of ABR algorithms and hence the Quality of Experience (QoE) of HAS users in Video on Demand (VoD) settings. Our results show that QUIC-level throughput data can indeed aid ABR algorithms to more accurately predict playout buffer under-runs, which in turn allows the ABR logic to take reactive measures in a timely fashion such that playback stalls can be avoided under challenging network conditions. Overall, our work presents a step towards improving ABR operation via cross-layer data exchange
HTTP/3's Extensible Prioritization Scheme in the Wild
No full textFor HTTP/2 and HTTP/3, multiple (Web page) resources are loaded by multiplexing them onto a single TCP or QUIC connection. A "prioritization system" is used to properly schedule the order in which the resources are sent. As HTTP/2's "prioritization tree" underperformed, a more straightforward setup called the Extensible Prioritization Scheme (EPS) was proposed for HTTP/3. This paper represents the first real-world measurement study into how this new scheme is supported and employed in practice by the three main browser engines and 12 different popular servers and cloud/CDN deployments. We find considerable heterogeneity in overall EPS (sub)feature support and even fundamental differences in approach/philosophy between the stacks. As incorrect prioritization can have a negative effect on (Web) performance metrics, our work not only provides essential insights for browser vendors and server deployments but also offers recommendations for future improvements
SLF4Web - MPEG-DASH datasets of static light fields
No full textMPEG-DASH datasets for the SLF4Web research project. SLF4Web is a Web-based implementation of a static light field consumption system; it allows SLF datasets to be adaptively streamed over the network (via MPEG-DASH) and then to be visualized in a vanilla Web browser. The datasets are encoded using the H.264/AVC video codec. A subset of the datasets are available in multiple qualities to allow for adaptive network streaming.
The SLF4Web source code is available on GitHub (https://github.com/EDM-Research/SLF4Web) and as a bundle at https://zenodo.org/badge/latestdoi/432214902
- …
