TY - JOUR AU - Nur, Afifah AU - Nugroho, Hendro PY - 2022/11/06 Y2 - 2024/03/29 TI - Detection Of Indonesian Sign System Letters (SIBI) Using Chain Code Method JF - POSITIF : Jurnal Sistem dan Teknologi Informasi JA - PJSTI VL - 8 IS - 1 SE - Articles DO - 10.31961/positif.v8i1.1133 UR - https://ejurnal.poliban.ac.id/index.php/Positif/article/view/1133 SP - 36 - 40 AB - <p><em>SIBI which stands for </em><em>Sistem Isyarat Bahasa Indonesia refers to an Indonesian signal language system for deaf and mute people. This research concerns with detecting the feature extraction of SIBI letters. For this reason, the researcher conducted several phases such as pre-processing, edge detection, image extraction, and letter similarity. In the segmentation process, manhattan distance method was carried out and then continued by converting RGB image to grayscale image and binary image. The next process namely mathematical morphology aimed at reducing the noise of image, whereas the method of chain code with eight directions of connectivity was employed as the extraction method of shape feature to determine the image probability. After that, the formation of eight connectivity of chain code in which edge was preciously detected by Robert operator generated probability value. Meanwhile, the euclidean distance method served as the equation of resulted probability value. By using 171 dataset consisting of 119 data reference and 52 data testing, the accuracy gained averagely 91%.</em></p> ER -