Implementation and Advanced Results on the Non-interrupted Skeletonization Algorithm

A modified K3M thinning algorithm

“…Three experiments were conducted using the original and improved K3M solutions as well as the KMM (Saeed et al, 2001) and Zhang-Suen (Zhang and Suen, 1984) approaches as reference algorithms. These experiments include:…”

“…The next one contains the description of the original K3M approach and its drawbacks. Afterwards, the proposed changes are described, followed by the results of experiments and a comparison of the improved approach with the original one as well as with the KMM (Saeed et al, 2001) and Zhang-Suen (Zhang and Suen, 1984) approaches. The conclusions are contained in the last section.…”

Intelligent Feature Extract System for Cursive-Script Recognition

“…Besides, the amount of the generated data will influence the speed of classification, as well. The contribution of the authors to this task is seen in [1,2,3].…”

“…In our work whenever needed we use KMM algorithm [2] described in details in [3,6] -writing it we paid particular attention to avoid the troubles with thick lines or rough edges of characters. Primary advantage of this method is that it gives a continuous one-pixel width skeleton for each line in a script.…”

“…It consists of projecting each pixel of a character onto axes forming a grid, in which the character is inscribed in its thinned form. The thinning procedure was performed by the mentioned above KMM algorithm [2,3]. The first versions of our projection method made use of two or four projection axes ( Fig.…”