Decomposition of 3D Objects into Geometric Primitives

Sakshi Cholli, Shrusti Girmath, Anilkumar Kulkarni

2025

Abstract

For applications in robotics, CAD systems, and scene comprehension, 3D objects must be broken down into their component geometric primitives. Current methods frequently depend on manually created features or regularised transformations, such as voxelization, which result in quantisation artifacts and inefficiencies. This study, which draws inspiration from PointNet, suggests a unified neural network architecture that breaks down 3D objects into basic geometric forms like spheres, cylinders, and planes by directly processing raw point clouds. Our model incorporates extra modules to learn local geometric characteristics for accurate decomposition, while utilizing PointNet’s capability to handle unordered point sets, guaranteeing permutation invariance.

Paper Citation

in Harvard Style

Cholli S., Girmath S. and Kulkarni A. (2025). Decomposition of 3D Objects into Geometric Primitives. In Proceedings of the 3rd International Conference on Futuristic Technology - Volume 3: INCOFT; ISBN 978-989-758-763-4, SciTePress, pages 496-502. DOI: 10.5220/0013622800004664

in Bibtex Style

@conference{incoft25,
author={Sakshi Cholli and Shrusti Girmath and Anilkumar Kulkarni},
title={Decomposition of 3D Objects into Geometric Primitives},
booktitle={Proceedings of the 3rd International Conference on Futuristic Technology - Volume 3: INCOFT},
year={2025},
pages={496-502},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013622800004664},
isbn={978-989-758-763-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Conference on Futuristic Technology - Volume 3: INCOFT
TI - Decomposition of 3D Objects into Geometric Primitives
SN - 978-989-758-763-4
AU - Cholli S.
AU - Girmath S.
AU - Kulkarni A.
PY - 2025
SP - 496
EP - 502
DO - 10.5220/0013622800004664
PB - SciTePress