Efficient parallel algorithms for the computation of two-dimensional image moments

Moments constitute an important set of parameters for image analysis. The low order moments contain significant information about a simple object. They have been used in finding the location and orientation of an object. Moment invariants have been used as features for pattern recognition. To compute moments of a two-dimensional image, a large number of multiplications and additions are required in a direct approach. Multiplications, which are the most time-consuming operations in simple processors, can be completely avoided in the proposed algorithms for low order moments. In this paper, parallel algorithms are proposed for efficient implementation in processor arrays. The basic idea is to decompose a 2-D moment into many vertical moments and a horizontal moment and to use the data parallelism for the vertical moments and the task parallelism for the horizontal moment.