Irawan, Hendra (2015) Rancang bangun sistem pengenalan huruf dan angka dalam SIBI (sistem isarat bahasa Indonesia) berbasis hand pose gesture recognition menggunakan microsoft kinect. Undergraduate thesis, Institut Teknologi Sepuluh Nopember Surabaya.
Preview |
Text
2411100089-Undergraduate_Thesis.pdf Download (8MB) | Preview |
Abstract
Telah dibangun sistem pengenalan huruf dan angka dalam SIBI (Sistem Isyarat Bahasa Indonesia) berbasis Hand Pose Gesture Recognition dengan menggunakan Microsoft Kinect sebagai solusi untuk memudahkan penderita tunarungu dan tunawicara dalam berkomunikasi dengan orang lain disekitarnya. Komponen utama yang digunakan dalam penelitian ini adalah Microsoft Kinect. Tangan-tangan dan jari-jari dideteksi dengan melakukan Segmentasi, K-means clustering, Graham Scan algorithm dan Moire contour tracing, identitas jari-jari dikenali dengan menghitung semua sudut antar jari dan ditentukan nilai ambang batasnya, sedangkan bahasa isyarat dikenali dengan algoritma Decision Tree Classifier. Penelitian ini diakhiri dengan karakterisasi sistem pengenalan bahasa isyarat. Sistem pengenalan yang dirancang dan dibangun dapat menerjemahkan huruf dan angka dalam SIBI dengan jarak terjauh yang dapat diterjemahkan sejauh 1.187,6 mm dan jarak terjauh mendeteksi tangan sejauh 1.438 mm dengan kecepatan komputasi sebesar 3,905 ms, dan dengan tingkat akurasi terendah menerjemahkan bahasa isyarat sebesar 30% dan total success rate sebesar 83,75%. Sistem juga dapat menerjemahkan bahasa isyarat disetiap kondisi pencahayaan. Medan pandang yang dimiliki oleh sensor Microsoft Kinect sebesar 58 derajat horizontal dan 45 derajat vertikal.
===================================================================================================
Letters and numbers recognition System in SIBI (Sistem Isyarat Bahasa Indonesia) base on Hand Pose Gesture Recognition has been built using Microsoft Kinect as a solution to make it easier for deaf and speech impaired in comunicating with other people around them. The main component used in this research is the Microsoft Kinect sensor. The hands and fingers are detected by doing a segmentation, K –means clustering, Graham Scan algorithm and Moire contour tracing. Identity of the fingers are known by calculating all of the angle between the finger and threshold limit values were specified, while sign language recognized by Decision Tree Classifier algorithm. The research concludes with a characterization of the sign language recognition system. The recognition system designed and built can translate alphabets and numbers in SIBI with the farthest distance which system can be translated is 1,187 .6 mm and the farthest distance to detect hands is 1.438 mm with 3,905 ms computational time, and the lowest accuracy of the system which can translate a sign language is 30% and the total success rate of the system is 83,75%. The system can also translate sign language in every lighting condition. Field of view owned by Microsoft Kinect sensor are 58 degrees horizontally and 45 degrees vertically.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSF 005.269 Ira r |
| Uncontrolled Keywords: | SIBI, Microsoft Kinect, Gesture Recognition, Hand Detection |
| Subjects: | T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques |
| Divisions: | Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis |
| Depositing User: | - Taufiq Rahmanu |
| Date Deposited: | 20 Nov 2019 06:55 |
| Last Modified: | 20 Nov 2019 06:55 |
| URI: | http://repository.its.ac.id/id/eprint/71915 |
Actions (login required)
 |
View Item |