Translasi Bahasa Isyarat

Rakhman, Juniar and Ramadijanti, Nana and Satriyanto, Edi (2010) Translasi Bahasa Isyarat. EEPIS Final Project.

Preview

PDF - Published Version Download (3169Kb) | Preview

Abstract

Due to their disability, hearing-impaired uses sign language as their primary means of communication. Sign language uses hand shape, facial expression, and body gesture. However this type of communication is unfamiliar to most people therefore communication between hearing-impaired and normal people can be difficult. There is a need for a system to translate sign language into spoken/written language, so that communication between hearing impaired and normal people can be simplified. To build this final project, the use of web camera is necessary to capture the image of the user�s hand. In general this program works by tracking the hand of the user, extract the shape of the hand, and then classify it. To track the position of the hand, we use HaarClassifier that has been trained prior to this project. To extract the hand we use skin detection and noise removal in which the resultan image will then be thresholded and normalized. After the image of the hand shape is extracted, the next step is to classify it by using k nearest neighbor algoritm. Each of the normalized binary images is then converted to feature vector, from this feature vector, distance is measured to find out the majority neighbor in which the new data belongs to. This system able to identify 19 fingerspelling signs out of 26 inteded signs to be identified. The average accuracy for this system is 89.68%. This value can vary depending on the data training consistency dan noise produced.

Actions (login required)