FEATURE-PRESERVING THINNING ALGORITHM FOR OPTICAL CHARACTER-RECOGNITION

We propose an improved thinning algorithm. Basically, the algorithm us es a 3 x 3 window to accommodate an eight-neighbor skeleton in each th inning iteration. This algorithm overcomes many thinning problems, suc h as Y-shaped distortions, spiky skeletons and skeleton shortening, an d thus preserves precise features of digital character patterns. By us ing this thinning algorithm, better structural features of a character pattern can be provided to an optical character recognition system, s uch that accurate recognition results can be achieved.