The two-dimensional (2D) and three-dimensional (3D) orthogonal moments are useful tools for 2D and 3D object recognition and image analysis. However, the problem of computation of orthogonal moments has not been well solved because there exist few algorithms that can efficiently reduce the computati
Two new algorithms for efficient computation of Legendre moments
β Scribed by J.D. Zhou; H.Z. Shu; L.M. Luo; W.X. Yu
- Publisher
- Elsevier Science
- Year
- 2002
- Tongue
- English
- Weight
- 147 KB
- Volume
- 35
- Category
- Article
- ISSN
- 0031-3203
No coin nor oath required. For personal study only.
β¦ Synopsis
Orthogonal moments have been successfully used in the ΓΏeld of pattern recognition and image analysis. However, the direct computation of orthogonal moments is very expensive. In this paper, we present two new algorithms for fast computing the two-dimensional (2D) Legendre moments. The ΓΏrst algorithm consists of transforming the pixel-based calculation of Legendre moments into the line-segment-based calculation. After all line-segment moments have been calculated, Hatamian's ΓΏlter method is extended to calculate the one-dimensional Legendre moments. The second algorithm is directly based on the double integral formulation. The 2D shape is considered as a continuous region and the contribution of the boundary points is used for fast calculation of shape moments. The numerical results show that the new algorithms can decrease the computational complexity tremendously, furthermore, they can be used to treat any complicated objects.
π SIMILAR VOLUMES
This paper presents a new algorithm for fast and accurate computation of Legendre moments. For a binary image, by use of a Green's theorem, we transform a surface integral to a simple integration along the boundary. The inter-order relationship of Legendre moments is then investigated. As a result,