Presentation

<
AGILE: Hand-object Interaction Reconstruction from Video via Agentic Generation
SessionHumans & Hands - Reconstruction
DescriptionAGILE reconstructs 3D hand-object interactions from monocular videos via agentic generation. It produces complete, simulation-ready object meshes and robustly estimates object pose without SfM, enabling accurate and stable results on challenging real-world data.
Authors
Jin-Chuan Shi
Zhejiang University
Binhong Ye
Zhejiang University
Tao Liu
Zhejiang University
Xiaoyang Liu
Zhejiang University
Yangjinhui Xu
Zhejiang University
Junzhe He
Zhejiang University
Zeju Li
Zhejiang University
Hao Chen
Zhejiang University
Chunhua Shen
Zhejiang University
Zhejiang University of Technology
Event Type
Technical Paper
TimeMonday, 20 July 20264:15pm - 4:25pm PDT
LocationRoom 403 A
Digital Library PDF
Session Time & Location
Sunday, 19 July 20266:00pm - 8:45pm PDTHall K
Monday, 20 July 20263:45pm - 5:35pm PDTRoom 403 A
Keywords
Generative AI
Geometry
Modeling
Registration Categories
Full Conference Supporter
Full Conference
Next PresentationNext PresentationEgoForce: Forearm-Guided Camera-Space 3D Hand Pose from a Monocular Egocentric Camera
Similar Presentations
Emerging Technologies Demo - AGILE: Hand-object Interaction Reconstruction from Video via Agentic Generation
LongE2V: Long-Horizon Event-based Video Reconstruction, Prediction, and Frame Interpolation with Video Diffusion Models
Composing People Together: Iterative Pose-Image Generation for Multi-Person Interaction Scenes