Jul 2026· Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Courses· pp. 1-5· 0 citations· 6 references
Abstract
This course explores how Universal Scene Description (USD) is being adopted in production as a foundation for modular assets, layered scene composition, artist-facing debugging, controlled edit flow, asset resolution, and rendering. Bringing together production examples from Pixar, Disney Animation, Autodesk, and LAIKA, the course emphasizes practical decisions and trade-offs rather than a single prescriptive pipeline. Beginning with USD foundations and the LIVERPS composition model, instructors explain how modern pipelines move from ad hoc and structured workflows toward modular, continuously composed systems. Subsequent sections focus on contextual debugging tools that make composition behavior understandable for artists, template-driven asset construction and deterministic resolution, rule-based edit forwarding and composition analyzers, and the redesign of asset structures for active production. Attendees will leave with a mental model for how USD structure affects scene behavior, how authored opinions resolve across layers, and how production teams can build, validate, and troubleshoot scalable USD assets and scenes.
Across scene-level 3D object detection, object pose estimation, and scene reconstruction, Lucida improves mAP over Boxer by 69% on R2S-Scene, raises ADD-SB@0.05 from 57.8% to 83.4% on CA-1M, and increases scene F-Score from 0.794 for SAM3D to 0.924.
Ming-Han Qin, Yuang Wang, Xiuyu Yang et al.· 1 citation· ⚡1
Frontier coding agents can now write and execute code that authors 3D environments, but whether they reliably understand 3D structure and precisely control scene state remains unclear. The generated 3D scene is a persistent, executable artifact: a convincing render can hide incorrect spatial relations, intersecting obj...
Xiao-Kang Ye, Siddhant Hitesh Mantri, Zi-Meng Chen et al.· 0 citations
Multimodal coding agents are expected to turn visual inputs into usable artifacts, and they act through a harness, the layer of tools, context management, and execution environment around the model. Existing evaluations often isolate short tool calls, API traces, or screenshot resemblance, and a low score under these p...
Li-Yang Fan, Chi Wei, Yi-Tai Li et al.· 0 citations
Structured-language models such as SceneScript reconstruct a scene as a short program of parametric commands, an inherently editable and semantically explicit representation. We ask three questions that stand between such models and their most compelling application, automated ingestion of existing buildings into BIM t...
P. Naikade, Thomas B. Moeslund, Andreas Møgelmose· 0 citations
Creating diverse physical simulations remains labor-intensive because assets, layout, physical parameters, motion, control, and rendering must be designed and debugged jointly. We present Text2Sim, a simulation-specialized agentic pipeline that converts a text-only request into an executable, editable dynamic case. Bui...
Xiao-Yu Xiong, T. Wang, Yi-Ling Qiao et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
