Loading...

An Automatic Persian Sign Language Recognition System Using Sensory Glove

Habibipour, Kamyar | 2013

1164 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 45649 (58)
  4. University: Sharif University of Technology, International Campus, Kish Island
  5. Department: Science and Engineering
  6. Advisor(s): Vossoughi, Gholamreza; Shamsollahi, Mohammad Bagher
  7. Abstract:
  8. Sign language is the main medium of communication for mute and deaf people. Unfortunately most hearing people do not understand this language. This fact causes communication difficulties for hearing impaired people and negatively affects their social life. This problem motivates researchers to develop speaking aids and helps deaf people to communicate with hearing people. Sign language includes concurrent combination of hand shapes, orientation and movements of the hands, arms or body, and facial expressions which make the character and word recognition a challenging task. Generally there are two fields of study in sign language recognition based on measuring devices: 1- camera based 2- Glove and sensor based. In the second approach user needs to wear cumbersome gloves but this method is usually more accurate and can help designers with portability of the device. Generally there are three major methods to solve sign language recognition problem: 1- Neural Network. 2- Hidden Markov Model. 3- Dynamic Programming Matching.
    In the present work, “Hidden Markov Model” which is a powerful statistical model has been utilized. Input devices include a sensory glove with 12 sensors which take care of hand posture data and IMU sensors mounted on wrist and arm which are responsible for capturing hand motion and orientation data. For sign recognition in isolated mode the proposed system is capable of recognizing 50 signs with recognition rate of 90 percent. For continuous mode (sentence recognition) recognition rate of the proposed system is 70 percent and there is no limitation imposed on context or number of signs in sentences.
  9. Keywords:
  10. Hidden Markov Model ; Sentence Recognition ; Sensory Glove ; Persian Sign Language ; Inertial Measurement Unite (IMU) ; Continuous Recognition

 Digital Object List

 Bookmark

No TOC