Self-organizing network modelling of 3D objects
Open Access
- Author:
- Liu, Runsang
- Graduate Program:
- Industrial Engineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- April 06, 2020
- Committee Members:
- Hui Yang, Thesis Advisor/Co-Advisor
Soundar Kumara, Committee Member
Robert Carl Voigt, Program Head/Chair
Lingzhou Xue, Committee Member - Keywords:
- Computer-aided design
Shape matching
Network modelling
object recognition - Abstract:
- Advanced manufacturing is moving towards a new paradigm of low-volume and high-mix production, e.g., the emergence of new additive manufacturing technology for layer-upon-layer fabrication of customized designs. There is an urgent need to develop effective representations of real-world 3D objects and further enable the matching and retrieval of engineering designs in a searchable database. Traditional mathematical representations of 3D objects, such as spherical harmonics descriptor, tends to encounter practical issues such as the uncertainty of voxelization due to the minute variations in mass center or object pose, and the use of L2 norm to derive the spectrum. This paper presents a new self-organizing network representation of 3D objects. Each voxel of the 3D object is a node in a network, and the edge is dependent on node closeness in space. Then, the network is self-organized by treating each node as an electrically charged particle, and further organizing the nodes with both attractive and repulsive forces. Experimental results show the effectiveness of network representation by reassembling the geometry of 3D objects. This network representation shows strong potential to improve the process of design automation for smart manufacturing.