Impact of Spectral Parameters on 3D Mesh Segmentation

Document Type : Original Article

Authors

1 Computer Science Department, Faculty of Computer and Information, Suez Canal university, Ismailia , Egypt.

2 Computer Science and Artificial Intelligence Department, College of Computer Science and Artificial Intelligence, University of Jeddah, Jeddah , Saudi Arabia

Abstract

A huge amount of research work has been devoted in recent years to spectral segmentation of 3D meshes composed of planar triangular faces. In this article, a thorough experimental investigation of three key spectral parameters is carried out to asses their impact on spectral segmentation. These parameters are the concavity threshold, c, the new segment size factor threshold, 𝜈 and the number of eigenvectors, K, of the affinity. For each parameter, experiments have been carried out to get an idea of the optimal value of the parameter, from the perspective of both the segmentation quality and the computational efficiency. The general conclusion of this work is that the optimal values of these parameters are model-dependent. The underlying causes for what properties in the model specify the optimal value of the parameter need deeply involved theoretical analysis, for which the present work has paved the way and provided intuitive pointers to consider.

Keywords

Main Subjects

Files

Share

How to cite

Statistics