Topology preserving parallel thinning algorithms

Abstract

Thinning is an iterative object reduction technique for extracting medial curves from binary objects. During a thinning process, some border points that satisfy certain topological and geometric constraints are deleted in iteration steps. Parallel thinning algorithms are composed of parallel reduction operators that delete a set of object points simultaneously. This article presents 21 parallel thinning algorithms for (8,4) binary pictures that are derived from the sufficient conditions for topology preservation accommodated to the three parallel thinning approaches.

Character Recognition Based on Skeletonization: A Survey.

“…Voronoi diagram 3. Thinning Skeletonization having the two main approaches; iterative and non-iterative [14]. As researches about both skeleton processing and adaptive binarization reach the final, conclusive state it became clear that the two preprocessing steps could be improved not only as separate entities but also as interacting stages of the document's image preparation [15].…”

An Accelerated Architecture Based on GPU and Multi-Processor Design for Fingerprint Recognition

“…Also thinning algorithms reduces the data that represents Minutiae and make the treatment more effective and faster. In the scientific literature, there are many iterative methods for thinning including sequential [49] or parallel [3]. In this step, the most used window size is a 3X3 pixel.…”

Robust and Efficient Fully Parallel 2D Thinning Algorithm

“…In such applications the object represented as a 2D binary image is a simple mapping from integer coordinates in 2D digital space Z 2 to {0 (white), 1 (black)}. The black point clouds form the objects and the white point clouds form the background and cavities [1,2,3,4]. One cannot obtain true skeletons in such a 2D digital space due to limitations of discretization.…”

“…Initially Kong [11] proposed sufficient conditions for 2D and 3D thinning which are modified to combine them with parallel thinning approaches by Nemath, Kardosh and Palagyi [2]. Nemath and Palagyi [3] proposed topology preserving algorithms based on sufficient conditions for topology and shape preservation. These algorithms produce spurs but are having less impact on later processing steps of the application.…”

“…Almost all fully parallel algorithms and parallel algorithms with sub-iterations failed in satisfying all the characteristics of skeletons despite of variety of techniques engineered to obtain the closure approximations to the true skeletons. These include: using variety of reduction operator ranging from 3x3 to 5x5 [1,3,4,6,8], efficient computations for 5x5 neighbor in terms of 3x3 neighbors [8], restore or retain templates to overcome the problem of discontinuities [1,6], flag map techniques [9], an extra pass to thin parts of skeletons with two pixel wide [7], and noise elimination techniques [6,8]. It is evident that no fully parallel algorithm exists which produces qualitative skeletons using only 3X3 mask based reduction operator without considering the techniques like flag map or retain templates, to preserve continuity.…”