STUDI PERBANDINGAN ALGORITMA THINNING ZHANG-SUEN DAN ALGORITMA THINNING STENTIFORD PADA PENGENALAN POLA CITRA KARAKTER TULISAN TANGAN

Thinning is a method in image morfology that can extract object skeleton of an image. Thinning method reduces original image pixels until gets one pixel thickness object of the image. This method is reliable for handwriting pattern recognition because the object sceletons which is the result of thinning can be useful as the important information in feature extraction and classification. There are three main processess in this experiment,they are: pre-processing, processing and analysis. Pre-processing level includes grayscalling, binerization, opening and closing operation. Processing level applies Zhang-Suen Thinning Algorithm and Stentiford Thinning Algorithm to thins the object of images. Meanwhile, the analysis level is carried out by timing run parameter and observes results with Kappa Cohen reliability test. This study uses 36 image samples The experiments were carried out with two different treatments, the first treatment used complete pre-processing and the second treatment is done without complete pre-processing. The result of experiment to handwriting images showed that Zhang-suen Thinning Algorithm without complete preprocessing is faster in timing run parameter, it spent 1,19 seconds. Meanwhile, Stentiford Thinning Algorithm with complete pre-processing is better for observation result parameter using Kappa Cohen reliability test