SEGMENTASI BIBIR DENGAN MENGGUNAKAN METODE KOMBINASI FCM-SNAKE

• Main Authors: , MUHAMMAD SHOLEKAN, , Faridah, S.T., M.Sc.
• 格式: Theses and Dissertations NonPeerReviewed
• 出版: [Yogyakarta] : Universitas Gadjah Mada 2014
• 主題: ETD
• 在線閱讀: https://repository.ugm.ac.id/132140/; http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=72657

Communication is a human need to develope. One of communication types is nonverbal communication. Nonverbal communication can be done by using lips reading patterns. The recognition of lips shape pattern must start with separating the lips and the skins. This separation can be done by using image segmentation, which separates between the object and the background. Many methods have been used to this segmentation, one of them are FCM ( Fuzzy Clustering Method ) and snake. This research will develop several of image segmentation algorithms. Each algorithm using image segmentation operation, starting color transformation to FCM (Fuzzy Clustering Method) and active contour snake. Each algorithm will be tested on 15 secondary data images and 6 primary data images. The data image is taken from variation in lips conditions and room illuminations. The success rate of algorithm measured by the value of error segmentation that is obtained. Error segmentation is the difference of keypoint distance coordinate between the original image and the result of segmentation image. From the comparation of the error segmentation value, it can be showed that the combination method of FCM-snake is the best algorithm with error segmentation value 15±10 pixel on secondary data images and error segmentation value 16±14 pixel on primary data images. The best illumination on secondary data image is between 330-340 lux for red lips condition and between 180-190 lux for pale lips and mustache lips. The best illumination on primary data image is between 260-270 lux (bright lighting condition).