Hand sign language translation for human computer interaction

• Main Author: Kailasam Parvathy
• Other Authors: Yap Kim Hui
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::Signal processing
• Online Access: http://hdl.handle.net/10356/15806
• Institution: Nanyang Technological University

Human Computer Interaction is a widely researched topic today. A significant amount of research in this field focuses on face recognition, speech recognition and gesture recognition. This project aims to create a hand-sign language translation system that is independent of any additional devices such as gloves with sensors or markers. Various feature extraction techniques have been evaluated in this project to determine the most suitable technique for the application in question. The feature extraction techniques evaluated were Principal Component Analysis, Linear Discriminant Analysis, Statistical Moments, Shape Signature, Elastic Bunch Graph Matching, Fourier Descriptors and Kernel PCA. The hand sign database used consists of 120 grayscale images (160 by 120 pixels) of 24 hand postures against a uniform background. 4 sets (96 images) will be used for training and cross-validation purposes and the remaining will be used in the testing process. The algorithms were evaluated on the basis of recognition rate and computation time. The best recognition rate was achieved for EBGM at 70% but with a computational time too long for use in a real-time application. The Fourier Descriptor method is very fast and has been recorded to achieve recognition rates as high as 94%. These methods are very robust in terms of translation, rotation and scale variations. The system implements a Graphical User Interface as well for the convenience of users.