Skin-Tone Segmentation in Real-Time Vision of Basic Hand Sign Through Pattern Recognition Using Hidden Markov

Authors

  • Raphael Baligod Malayan Colleges Laguna, Cabuyao City, Laguna 4025 Philippines.
  • Leis Alemania Malayan Colleges Laguna, Cabuyao City, Laguna 4025 Philippines.
  • Alaiza P. Dabatos Malayan Colleges Laguna, Cabuyao City, Laguna 4025 Philippines.
  • Adomar L. Ilao Malayan Colleges Laguna, Cabuyao City, Laguna 4025 Philippines.

Keywords:

Hand Sign, Skin Color, Pattern Recognition, Hidden Markov Model, Haar Classifier,

Abstract

The study applied computer vision to perform hand sign recognition commonly affected by skin color during image segmentation. Through YCbCr color algorithm, image color which is based on RGB color space was converted to pure color space which separates luminance and chrominance components resolving skin tone issue of extracting hand contour from real-time right-hand sign. A prototype was developed to recognized hand sign captured by a low-cost web camera. The captured hand sign image passed through canny edge detection and absolute difference threshold technique to construct 2D hand contour in conjunction with Haar classifier and Hidden Markov Model in providing a real-time American Sign Language (ASL) interpretation of the demonstrated hand sign.

Downloads

Download data is not yet available.

Downloads

Published

2017-06-01

How to Cite

Baligod, R., Alemania, L., P. Dabatos, A., & L. Ilao, A. (2017). Skin-Tone Segmentation in Real-Time Vision of Basic Hand Sign Through Pattern Recognition Using Hidden Markov. Journal of Telecommunication, Electronic and Computer Engineering (JTEC), 9(2-3), 145–148. Retrieved from https://jtec.utem.edu.my/jtec/article/view/2344