SimpModeling: Sketching Implicit Field to Guide Mesh Modeling for 3D Animalmorphic Head Design
summary
Two-stage sketch pipeline separates global shape and local detail: closed-contour input and curve handles produce M0 then deform to coarse M1 using a voxel-aligned implicit network, while surface strokes plus front/back depth and a pix2pixHD-predicted normal map drive a pixel-aligned occupancy field to carve M2. Trained on the new 3DAnimalHead dataset (largest animalmorphic head set reported) and tested with 15 amateur users, the system yields more controlled global geometry and richer local details than fiber-surface methods.
AI-written summary · about
chapters (8)
- 0:00Introduction to SimpModeling
- 1:31Sketch-based modeling history
- 3:37Two-stage modeling approach
- 5:04Workflow demonstration
- 6:29Coarse mesh modeling stage
- 7:11Geometric detail crafting stage
- 8:14Algorithm dynamic process
- 9:17Dataset and results overview
related talks
12mOutline and Detail: A Semantic-Driven Framework for Layered 2D Character Generation
22mDreamSketch: Early Stage 3D Design Explorations with Sketching and Generative Design
21mSketchingWithHands: 3D Sketching Handheld Products with First-Person Hand Posture
47m3D Digital Avatars with Machine Learning
14mAutocomplete Animated Sculpting
37mThe principle from the general to the particular - or how to work less and do more
19mSketchforme: Composing Sketched Scenes from Text Descriptions for Interactive Applications
20mLightning Talk: Beyond Polygons: The Future of 3D Graphics with Implicit Models