Sign Language Translator Application Using OpenCV

International Conference on Advanced Materials for Better Future 2017

Triyono, L.^1 ; Pratisto, E.H.^2 ; Bawono, S.A.T.^2 ; Purnomo, F.A.^2 ; Yudhanto, Y.^2 ; Raharjo, B.^2

Informatics Department, Politeknik Negeri Semarang, Jl. Prof. Sudharto, Tembalang, Semarang, Jawa Tengah

50275, Indonesia^1

Informatics Engineering Department, Sebelas Maret University, Jl. Ir. Sutami 36A, Kentingan Jebres Surakarta

57126, Indonesia^2

关键词: Hand gesture;    Hu moments;    Light intensity;    Outdoor environment;    Research focus;    Search method;    Sign language;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/333/1/012109/pdf
DOI  :  10.1088/1757-899X/333/1/012109

来源: IOP

【 摘 要 】

This research focuses on the development of sign language translator application using OpenCV Android based, this application is based on the difference in color. The author also utilizes Support Machine Learning to predict the label. Results of the research showed that the coordinates of the fingertip search methods can be used to recognize a hand gesture to the conditions contained open arms while to figure gesture with the hand clenched using search methods Hu Moments value. Fingertip methods more resilient in gesture recognition with a higher success rate is 95% on the distance variation is 35 cm and 55 cm and variations of light intensity of approximately 90 lux and 100 lux and light green background plain condition compared with the Hu Moments method with the same parameters and the percentage of success of 40%. While the background of outdoor environment applications still can not be used with a success rate of only 6 managed and the rest failed.