A fast algorithm for ranking and perimeter computation on a reconfigurable mesh computer

In this paper we present a parallel algorithm for solving the ranking problem in the connected components of images. This algorithm is assigned to be implemented on an n×n reconfigurable mesh computer, where the image is stored one pixel per processing element. This algorithm is achieved in three essential stages, such as contour extraction of the component, straight line segment identification in the contour and ranking the boundary pixels. This algorithm is of O(log p) times complexity, where p is the number of the straight line segments in the contour. The reconfigurability of the mesh allows us to perform all the mentioned procedures fast and independently of the image size