Eurographics Digital Library
Not a member yet
17121 research outputs found
Sort by
Distance-Aware Tri-Perspective View for Efficient 3D Perception in Autonomous Driving
No full textThree-dimensional environmental perception remains a critical bottleneck in autonomous driving, where existing vision-based dense representations face an intractable trade-off between spatial resolution and computational complexity. Current methods, including Bird's Eye View (BEV) and Tri-Perspective View (TPV), apply uniform perception precision across all spatial regions, disregarding the fundamental safety principle that near-field objects demand high-precision detection for collision avoidance while distant objects permit lower initial accuracy. This uniform treatment squanders computational resources and constrains real-time deployment. We introduce Distance-Aware Tri-Perspective View (DA-TPV), a novel framework that allocates computational resources proportional to operational risk. DA-TPV employs a hierarchical dual-plane architecture for each viewing direction: low-resolution planes capture global scene context while high-resolution planes deliver fine-grained perception within safety-critical reaction zones. Through distance-adaptive feature fusion, our method dynamically concentrates processing power where it most directly impacts vehicle safety. Extensive experiments on nuScenes demonstrate that DA-TPV matches or exceeds single high-resolution TPV performance while reducing memory consumption by 26.3% and achieving real-time inference. This work establishes distance-aware perception as a practical paradigm for deploying sophisticated three-dimensional understanding within automotive computational constraints. Code is available at https://github.com/yytang2012/DA-TPVFormer.Pacific Graphics Conference Papers, Posters, and DemosPoint Clouds & Gaussian Splattin
A Scan-to-BIM Approach for the Preservation of the Architectural and Archaeological Heritage: the Digitisation of the Complex of San Nicolò Regale in Mazara del Vallo (Italy)
No full textIn recent years, Built Heritage (BH) management has greatly benefited from advances in digital survey technologies and the application of Building Information Modelling (BIM) processes for digitisation and preservation purposes. The integration of range-based (static and mobile) and image-based (aerial and terrestrial) survey techniques enables the creation of comprehensive 3D geospatial datasets for BIM modelling. This study presents the Heritage BIM (HBIM) digitisation of an architectural and archaeological complex in Mazara del Vallo (Italy). The site includes two assets: the 12th-century Arab-Norman church of San Nicolò Regale and the remains of a late Roman domus (3rd-4th century C.E.) beneath it, featuring polychrome mosaic floors. This digitisation aimed to provide exhaustive and detailed geometric documentation and create an accurate digital model for conservation and management purposes. Data acquisition combined mobile and static laser scanning with close-range photogrammetry. The final model was obtained through a Scan-to-BIM approach for the church and an integrated Scan-to-Mesh-to-BIM approach for the domus, ensuring accurate representation of both architectural and archaeological elements. This methodology serves as a reference for digitising and preserving similar heritage sites.Digital HeritageAnalysing and Documenting Digitized Asset
Real-Time Rendering Framework for Holography
No full textWith the advent of holographic near-eye displays, the need for rendering algorithms that output holograms instead of color images emerged. These holograms usually encode phase maps that alter the phase of coherent light sources such that images result from diffraction effects. While common approaches rely on translating the output of traditional rendering systems to holograms in a post processing step, we instead developed a rendering system that can directly output a phase map to a Spatial Light Modulator (SLM). Our hardware-ray-traced sparse point distribution, and depth mapping enable rapid hologram generation, allowing for highquality time-multiplexed holography for real-time content. Additionally, our system is compatible with foveated rendering which enables further performance optimizations.Computer Graphics ForumReal-Time Rendering: Fast, Furious, and Accurate44
On-Device 3D Point Cloud Segmentation for Mixed Reality Applications
No full textThis paper deals with preliminary research on segmenting 3D space captured by the depth sensor integrated into the HTC Vive XR Elite headset. Obtained cloud of 3D points was employed as the basis for a segmentation algorithm that identifies individual objects by detecting planes and grouping them into bounding boxes. Since spatial data is not available in the socalled PCVR mode, all computations had to be performed directly on the head-mounted display (HMD) headset, which is heavily constrained in terms of performance. The proposed approach provides a practical foundation for replacing detected objects with virtual counterparts (overlaying), while meeting computational and energy efficiency requirements for prolonged headset operation.ICAT-EGVE 2025 - International Conference on Artificial Reality and Telexistence and Eurographics Symposium on Virtual Environments - Posters and DemosPosters and Demo
BASS-MLIC: a Novel Synthetic Dataset for Single-View Inverse Rendering Tasks on Cultural Heritage Artifacts
No full textBASS-MLIC is a synthetic dataset created with Blender to support multi-light image processing and inverse rendering tasks. It features orthographic views of culturally significant surfaces rendered with realistic materials and includes rich ground truth annotations, such as normals, depth, shadows, materials, and BRDF parameters. These annotations enable evaluation across diverse tasks like relighting, Photometric Stereo, shadow-aware estimations, and BRDF fitting. Preliminary experiments highlight its practical utility.Smart Tools and Applications in Graphics - Eurographics Italian Chapter ConferencePoster
Fluidly Revealing Information: A Survey of Un/foldable Data Visualizations
No full textRevealing relevant information on demand is an essential requirement for visual data exploration. In this state-of-the-art report, we review and classify techniques that are inspired by the physical metaphor of un/folding to reveal relevant information or, conversely, to reduce irrelevant information in data visualizations. Similar to focus+context approaches, un/foldable visualizations transform the visual data representation, often between different granularities, in an integrated manner while preserving the overall context. This typically involves switching between different visibility states of data elements or adjusting the graphical abstraction linked by gradual display transitions.We analyze a literature corpus of 101 visualization techniques specifically with respect to their use of the un/folding metaphor. In particular, we consider the type of data, the focus scope and the effect scope, the number of un/folding states, the transformation type, and the controllability and interaction directness of un/folding. The collection of un/foldables is available as an online catalog that includes classic focus+context, semantic zooming, and multiscale visualizations as well as contemporary un/foldable visualizations. From our literature analysis, we further extract families of un/folding techniques, summarize empirical findings to date, and identify promising research directions for un/foldable data visualization.Computer Graphics ForumNovel Approaches to Visualization Adaptabilit
IntelliCircos: A Data-driven and AI-powered Authoring Tool for Circos Plots
No full textGenomics data is essential in biological and medical domains, and bioinformatics analysts often manually create circos plots to analyze the data and extract valuable insights. However, creating circos plots is complex, as it requires careful design for multiple track attributes and positional relationships between them. Typically, analysts often seek inspiration from existing circos plots, and they have to iteratively adjust and refine the plot to achieve a satisfactory final design, making the process both tedious and time-intensive. To address these challenges, we propose IntelliCircos, an AI-powered interactive authoring tool that streamlines the process from initial visual design to the final implementation of circos plots. Specifically, we build a new dataset containing 4396 circos plots with corresponding annotations and configurations, which are extracted and labeled from published papers. With the dataset, we further identify track combination patterns, and utilize Large Language Model (LLM) to provide domain-specific design recommendations and configuration references to navigate the design of circos plots. We conduct a user study with 8 bioinformatics analysts to evaluate IntelliCircos, and the results demonstrate its usability and effectiveness in authoring circos plots.Computer Graphics ForumExplainable and Generative A
Towards Scalable Out-of-Core Volume Rendering for High-Performance Visualization of 3D Temporal Multivariate Gas and Fluid Data
No full textScientific visualization enables a deeper insight into the data structure and properties. For Gas and fluid simulations, particularly those resulting from real-world gas captures or Computational Fluid Dynamics (CFD) models, this requires the analysis and integration of multivariate attributes alongside dynamic temporal variations. We present an ongoing work to develop a scalable, Out-of-core volume rendering application designed for massive 3D temporal multivariate datasets. Our rendering framework, called FRIGAS (Fluid Rendering Infrastructure for Gas and Atmospheric Simulations), introduces a novel pipeline which leverages improvements in Direct Volume Rendering (DVR) techniques for large scale visualizations, addressing constraints regarding temporal structured data and multivariate analysis for interactive spatial and temporal navigation without HPC resources.EuroVis 2025 - PostersPoster
A Knowledge Graph for Andean Ritual Heritage: Structuring the Chunchada Dance in Honor of the Virgin of Rosario (Paucartambo, Peru)
No full textWe present a knowledge graph to document the Chunchada-a women-led ritual dance in honor of the Virgin of Rosario in Paucartambo, Peru. Using oral testimonies, archival data, and AI-based extraction, we reveal ritual relations through graph analysis and community curation. The project exemplifies a hybrid approach to intangible heritage preservation, combining graph mining and community-driven data curation to facilitate both academic insight and public engagement.Digital HeritagePredictive Analysis, AI, Simulation, and Novel Computational Method
Comparing VR and AR in Cultural Heritage Active Learning: A Study Based on the Stimulus-Organism-Response Model and the Engagement Theory
No full textWith the rapid advancement of immersive technologies, Virtual Reality (VR) and Augmented Reality (AR) have emerged as powerful tools for enhancing cultural heritage experiences. These technologies offer opportunities to engage learners through interactive and rich content, transforming how cultural knowledge is disseminated and understood. In this study, we developed two systems - one in VR with a head-mounted display and the other in AR with a mobile device for cultural heritage exploration. By employing a comprehensive survey with 70 responses, we explored the determinants of active learning within two systems. The results showed that the effects of interactivity and content richness on active learning are mediated by cognitive engagement toward both VR and AR systems. Our findings contribute valuable insights to the field of technology-mediated learning and provide practical guidelines for optimizing immersive cultural heritage experiences through targeted design strategies, highlighting the potential gaps in content design optimization.Digital HeritageModern Technologies for Serious Gaming in Cultural Heritag