A multiscale parallel thinning algorithm
This paper presents a new parallel thinning algorithm. Thinning algorithms based on contour tracing are strongly sequential and irregular, with a complexity of O(n2). Thinning algorithms using parallel neighboring operations are regular with a complexity of O(n3).
The parallel thinning approach we propose uses the notion of multi-scaling. This new algorithm keeps the regularity and parallelism of the neighboring operator while achieving a cost of O(n2). Complexity analysis and comparison with a well-known thinning algorithm using a SIMD machine are presented.
Unable to display preview. Download preview PDF.
- 1.T. Pavlidis. A thinning algorithm for discrete binary images. Computer vision and image processing, 20:142–157, 1980.Google Scholar
- 3.C.J. Hilditch. Comparison of thinning algorithms on a parallel processor. Computer Vision, Graphics and Image Processing, 1(3):115–132, 1983.Google Scholar
- 4.C. Arcelli. Pattern thinning by contour tracing. Comp. Graph. and Image Proc., 17:130–144, 1981.Google Scholar
- 6.C.M. Holt, A. Stewart, M. Clint, and R.H. Perrott. An improved parallel thinning algorithm. Comm. ACM, 30:156–160, 1987.Google Scholar
- 7.R.T. Chin, H.K. Wan, D.L. Stover, and R.D. Iverson. A one-pass thinning algorithm and its parallel implementation. Comp. Vision, Graph. and Image Proc., 40:30–40, 1987.Google Scholar
- 8.V. Poty and S. Ubeda. Parallel thinning algorithm using k×k mask. In Int. Coll. on Parallel Image Proc., 337–352, 1991.Google Scholar
- 9.J. Olszewski. A flexible thinning algorithm allowing parallel, sequentiel and distributed application. To appear in ACM Trans. on Mathematical Sofware.Google Scholar