BEGIN:VCALENDAR VERSION:2.0 PRODID:Linklings LLC BEGIN:VTIMEZONE TZID:America/Los_Angeles X-LIC-LOCATION:America/Los_Angeles BEGIN:DAYLIGHT TZOFFSETFROM:-0800 TZOFFSETTO:-0700 TZNAME:PDT DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:-0700 TZOFFSETTO:-0800 TZNAME:PST DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260417T190159Z LOCATION:West Building\, Level 2\, Outside Room 219 DTSTART;TZID=America/Los_Angeles:20250812T090000 DTEND;TZID=America/Los_Angeles:20250812T173000 UID:siggraph_SIGGRAPH 2025_sess434@linklings.com SUMMARY:Posters: Rendering DESCRIPTION:68. Evaluating Skin Tone Biases in Virtual Human Rendering\n\n We present a comparative analysis of skin tone rendering in MetaHuman avat ars using real-world and reference-based color inputs, revealing systemati c differences across the Monk Skin Tone scale and highlighting key limitat ions in current real-time rendering pipelines for darker and intermediate skin...\n\n\nErick Menezes (UNIT); Helena Leal (UNIT, UFS); João Vítor Mou ra (UNIT); Victor Araujo (PUCRS/INCT-SANI, UNIT-SE); and Soraia Raupp Muss e (PUCRS)\n---------------------\n75. To Infinity and Beyond: a GPU-Driven Memory Sharing Pipeline to Generate and Process Infinite Synthetic Data\n \nOur GPU-resident pipeline based on Unreal Engine 5 unifies scene generat ion, rendering, and processing entirely on the GPU, eliminating CPU–GPU tr ansfers and disk I/O to achieve near-constant per-sample latency, up to 12 × speedups, and sustained high-throughput training with effectively infini ...\n\n\nDaniele Della Pietra, Gino Lanzo Hahn, and Nicola Garau (Universi ty of Trento)\n---------------------\n74. SurfelPlus: A Surfel-Based Globa l Illumination Solution Optimized for Low-End Graphics Hardware\n\nSurfelP lus introduces a real-time global illumination renderer optimized for low- end hardware, achieving dynamic indirect lighting through unified surfel g eneration, adaptive surfel management, and advanced spatial-temporal filte ring, significantly improving performance and visual fidelity without e... \n\n\nRuipeng Wang, Zhen Ren, and Jinxiang Wang (University of Pennsylvani a)\n---------------------\n73. Scalable Volume Rendering of Billion-Cell C FD Simulations Using VFX Pipelines\n\nWe present a high-performance, VFX-i nspired workflow that transforms unstructured CFD data into scalable, high -fidelity visualizations using parallel voxelization, OpenVDB export, and CyclesPhi rendering, supporting both batch processing and interactive fram e exploration for scientific analysis and ...\n\n\nPetr Strakos, Milan Jar os, Tomas Brzobohaty, Marketa Faltynkova, Ondrej Meca, and Lubomir Riha (I T4Innovations, VSB - Technical University of Ostrava)\n------------------- --\n76. Towards Accelerating Polarization Path tracing of Multi-Bounce Smi th Microfacet BSDFs\n\nThis paper proposes a polarization path tracing met hod that incorporates multiple microfacet reflections and introduces an ap proximation potentially enabling computationally efficient rendering.\n\nI NVITED TO THE FIRST ROUND OF THE STUDENT RESEARCH COMPETITION\n\n\nHidehit o Ohba (Waseda University), Tatsuya Yatagawa (Hitotsubashi University), an d Shigeo Morishima (Waseda University)\n---------------------\n72. Interac tive Object Insertion With Differentiable Rendering\n\nWe develop an objec t insertion pipeline and interface that enables iterative editing of illum ination-aware composite images. Our pipeline leverages off-the-shelf compu ter vision methods and differentiable rendering to reconstruct a 3D repres entation of a given scene. Users can add 3D objects and ren...\n\n\nWeikun Peng, Sota Taira, Chris Careaga, and Yağız Aksoy (Simon Fraser University )\n---------------------\n70. Hide A Bit: A Training-Free and High-Fidelit y Steganography Method for 3D Gaussian Splatting Based on Bit Manipulation and RSA Encryption\n\nThis paper presents a high-fidelity steganography m ethod for 3D Gaussian splatting that requires no additional training. We p ropose a bit-level embedding technique that leverages the lower bits of th e 32-bit floating-point representation. We further introduce an embedding strategy that considers eac...\n\n\nKaoru Sasaki, Kazuhito Sato, and Shugo Yamaguchi (Waseda University) and Keitaro Tanaka and Shigeo Morishima (Wa seda Research Institute for Science and Engineering)\n-------------------- -\n69. GBake: Baking 3D Gaussian Splats Into Reflection Probes\n\nGBake in troduces a raytracing-based technique for creating reflection probes in Ga ussian-splatted environments that overcomes inherent EWA splatting errors at cubemap seams, enabling realistic integration of traditional mesh objec ts into scenes comprised of 3D Gaussians.\n\nINVITED TO THE FIRST ROUND .. .\n\n\nStephen Pasch (Columbia University); Joel Salzman (Columbia Univers ity, Brown University); and Changxi Zheng (Columbia University)\n--------- ------------\n71. HyperParamBRDF: Fast Parametric Reflectance via Hypernet works and Physics-Based Simulation\n\nHyperParamBRDF uses hypernetworks co nditioned on physical parameters to predict nanostructure BRDFs with high fidelity, accelerating appearance evaluation by orders of magnitude compar ed to simulation and enabling real-time exploration.\n\n\nAbraham Beauferr is and Wei Sen Loi (Huawei Canada)\n---------------------\n67. Efficient P roxy Raytracer for Optical Systems Using Implicit Neural Representations\n \nRay tracing is a widely used technique for modeling optical systems, inv olving \nsequential surface-by-surface computations which can be computati onally \nintensive. We propose Ray2Ray, a novel method that leverages impl icit neural representations to \nmodel ray tracing through optical systems with gre...\n\n\nShiva Sinaei (University of Osaka), Chuanjun Zheng and K aan Akşit (University College London (UCL)), and Daisuke Iwai (University of Osaka)\n\nRegistration Category: Full Conference, Experience END:VEVENT END:VCALENDAR