Sign in to view source links and access this dataset
Description
24 sequences simulate articulated object motion for 17 categories from PartNet-Mobility, including Camera, Chair, and Refrigerator. The dataset supports research on reconstructing articulated digital twins from monocular video, as presented in the associated paper by author ShawnRicardo.
Use Cases
Train models to reconstruct articulated 3D geometry and motion from monocular video using the 17 object categories.
Benchmark kinematic estimation algorithms on sequences featuring complex kinematics from categories like Door, Oven, and Eyeglasses.
Develop category-specific motion priors for articulated objects such as Chair, Table, and Storage Furniture.
Strengths
Covers 17 distinct articulated object categories from the PartNet-Mobility benchmark.
Contains 24 sequences designed to feature complex kinematic motion.
Created for a published research paper on articulated 3D reconstruction.
Limitations
Dataset size, row count, and specific file formats are not provided in the input.
Limited to simulated sequences, which may not capture the full complexity of real-world video data.
The 17 categories represent a specific subset of objects, potentially limiting generalizability.
Provenance
Source
Official repository for the paper 'Articulat3D: Reconstructing Articulated Digital Twins From Monocular Videos with Geometric and Motion Constraints'.
Collection Method
Simulated sequences based on 17 object categories from PartNet-Mobility.
Time Range
null
Freshness
Last updated March 2026.
Geography
null
The full description is hosted externally; users should review the dataset page on Hugging Face for complete details on access and structure.