Scale Invariants from Gaussian-Geometric Moments
Tao Sun, Shulin Yang, Bin Wu · Proceedings of the 2021 5th International Conference on Electronic Information Technology and Computer Engineering · 2021
The Gaussian geometric moment is obtained by adding a Gaussian kernel to the geometric moment, which has better characteristic expression ability and robustness than the geometric moment. Based on the construction theory of geometric moment invariants, the translation and rotation invariants of Gaussian geometric moments can be easily constructed. However, due to the existence of the Gaussian kernel, the method of constructing scale invariants in geometric moments is no longer applicable. In this paper, by introducing the variable factor σ, the Gaussian geometric moment invariants for scale transformation is constructed, and the construction method of this scale invariants will not affect the translation invariance and rotation invariance. Numerical experiments verify the validity of scale invariants, and at the same time have better numerical stability than geometric moments.