DocumentCode :
1732766
Title :
Measurement of Motion Quantity from human movement and detection of the sign language word
Author :
Takai, Miwa
Author_Institution :
Dept. of Comput. Intell. & Syst. Sci. Interdiscipl. Grad. Sch. of Sci. & Eng., Tokyo Inst. of Technol., Tokyo, Japan
fYear :
2012
Firstpage :
298
Lastpage :
303
Abstract :
This study measures Motion Quantity of the movement of the sign language speaker, and detects the characteristic posture of each sign language word from the sign language so that we understand the meaning of sentence of the sign language. The sign language consists of one or plural sign language word(s). Speaker takes the unique movement of each sign language word. The speaker puts these sign language words together as well as a sentence, and tells the meaning of sentence to listener of the sing language. The proposal method detects characteristic posture from the movement that a speaker takes in a sign language word by extracting characteristic posture with Motion Quantity. Motion Quantity finds the value based on the size of Motion Region occurring to human body from the movement that a speaker takes. Therefore, speed of the movement of the sign language varies according to the skill of the speaker, but this proposal method can extract characteristic posture from the quick movement of the speaker using Motion Quantity. In the experiment, the proposal method succeeded in detecting the characteristic posture of each sing language word to constitute the sign language of the speaker.
Keywords :
feature extraction; gesture recognition; image motion analysis; word processing; characteristic posture extraction; human movement; motion quantity measurement; motion region size; sign language speaker movement; sign language word detection; sing language listener; speaker skills; Manganese; Mechatronics; Motion measurement;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Advanced Mechatronic Systems (ICAMechS), 2012 International Conference on
Conference_Location :
Tokyo
ISSN :
1756-8412
Print_ISBN :
978-1-4673-1962-1
Type :
conf
Filename :
6329727
Link To Document :
بازگشت