Lune

IEEE VR2026Top-tier venue

GenerativeSurface: GenAI-Assisted Semantic Scaffolding for 3D Conceptual Design in Mixed Reality

Yichen Chai, Qianyuan Zou, Zihang Liu, Yi Wang, Henry Been-Lirn Duh

2026Year

Abstract

The transition from conceptual sketch to 3D model poses a major challenge in engineering design. While Mixed Reality (MR) enables direct 3D creation, designers face two key challenges: the Initiation Challenge (lack of meaningful starting points) and the Expression Challenge (difficulty in precise mid-air sketching). We introduce GenerativeSurface, leveraging AI to translate natural language descriptions into interactive 3D surfaces that aid designers in 3D sketching, simultaneously addressing both challenges. Our system transforms the design workflow into a fluid "describe, generate, sketch, iterate" cycle. Comparative evaluation against primitive scaffolding and freehand mid-air sketching demonstrates that our approach effectively bridges the gap between designers' conceptual intent and 3D sketch expression in immersive environments.

Ask about this paper

Ask your agent about it.

Lune has read the top-tier papers around this one, so every answer names the papers it rests on.

Questions to start from

Your agent calls

Lunesearch_papers

Ask in Lune

Free to start. No credit card required.

lune papers get c7020a0b-a418-424c-bc80-ae06baf718a8

Related papers

Dusk over the sea between two cliffs drawn in fine vertical lines