2026-09-21T21:46:02Z https://www.ideals.illinois.edu/oai-pmh

oai:www.ideals.illinois.edu:2142/125661 2026-02-03 col_2142_5131 col_2142_10761 com_2142_5130 com_2142_10755 com_2142_234

application/pdf en text Submission published under a 24 month embargo labeled 'U of I Access', the embargo will last until 2026-08-01 The student, Wei Jiang, accepted the attached license on 2024-05-23 at 00:39. The student, Wei Jiang, submitted this Thesis for approval on 2024-05-23 at 00:39. This Thesis was approved for publication on 2024-05-24 at 15:40. DSpace SAF Submission Ingestion Package generated from Vireo submission #20810 on 2025-02-04 at 21:15:39 Building rearticulable models for 3D articulated objects from multi-view RGB videos Jiang, Wei 2024-05-24 Wang, Shenlong 3d Articulated Object Understanding Gaussian Splatting Neural Rendering Part Discovery eng This thesis presents a solution to the novel task of building a 3D rearticulable models from multi-view colored (RGB) videos, which enables to render the articulable object at sampled state, discover the rigid parts, and reconstruct the kinematic structure. The solution is comprised of Gaussian Articulated Implicit Model (GAIM), a representation for dynamic scenes with an articulable object, and a relax-and-project-based optimization framework. Evaluated on the Watch-It-Move (WIM) dataset, the proposed solution achieves on par view synthesis performance as other rendering baselines, and can effectively discover rigid parts and properly reconstruct the kinematic model of the observed articulated object. 2024-08 Thesis https://hdl.handle.net/2142/125661 Copyright 2024 Wei Jiang Computer Science Computer Science University of Illinois at Urbana-Champaign M.S. Thesis