3D Gaussian Geometric Moment Invariants
Tao Sun,Tao SunSchool of Big Data and Basic Science,Shandong Institute of Petroleum and Chemical Technology,Dongying,Shandong,China
DOI: https://doi.org/10.1080/08839514.2024.2318983
IF: 2.777
2024-02-26
Applied Artificial Intelligence
Abstract:3D moment invariants are important tools for 3D image feature representation. In this paper, we introduced a novel approach for constructing 3D moment invariants using Gaussian geometric moments. Our proposed method demonstrated invariance under translation, rotation, and scale transformations. The numerical experiments validate the invariance and robustness of the proposed method, comparing it with traditional 3D geometric moments and revealing superior performance in the presence of noise and transformations. Additionally, the method is applied to content-based 3D image retrieval, exhibiting promising results through Minkowski distance-based retrieval on the Princeton Shape Benchmark (PSB) database.
computer science, artificial intelligence,engineering, electrical & electronic