A Portable and Real Time Taiwan Natural Sign Language Translation System

• Main Authors: Boxun - Xun Zhuo, 卓伯勳
• Other Authors: Lih-Jen Kau
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/2fwk7z

碩士 === 國立臺北科技大學 === 電子工程系研究所 === 105 === This paper presents a wearable data gloves combined with mobile devices. The establishment of sign language translation vocabulary library is mainly based on the natural sign language in northern Taiwan. We propose a sign language gesture recognition algorithm that divides the gesture of the corresponding vocabulary into hand, face, and hand movement (e.g., posture, orientation, and motion). Defining the degree of bending status on each finger by detecting the user’s finger bending action through flex sensors set in the gloves; analyzing the face of the palms and gesture movement through the accelerometer and the gyroscope of the inertial sensor and combine Finite-State Machine into compound of sign language vocabulary gesture. Categorizing the sign language gesture into static gesture words, dynamic gesture words and compound gesture words and arranging the sign language vocabulary to the sign language gesture dictionary in this search method. Sending the finger bending of the hand, face of the palms, hand movements and characteristics likewise to the mobile device database through blue tooth and make identifying comparison with the sign language gesture dictionary. Utilizing the vocal device of the mobile device, sending the corresponding vocal information to the speaker of the mobile device and showing the vocabulary on the screen of the mobile device can make deaf-mutes understand the message from the sign language even if they don’t understand the sign language.