Comparative Study of Segmentation Technique for Hand Gesture Recognition System

Authors

  • Farah Farhana Mod Ma'asum Centre for Electronics Engineering Studies (CEES), Faculty of Electrical Engineering, Universiti Teknologi Mara, 40450 Shah Alam, Selangor, Malaysia.
  • Suhana Sulaiman Centre for Electronics Engineering Studies (CEES), Faculty of Electrical Engineering, Universiti Teknologi Mara, 40450 Shah Alam, Selangor, Malaysia.
  • Azilah Saparon Centre for Electronics Engineering Studies (CEES), Faculty of Electrical Engineering, Universiti Teknologi Mara, 40450 Shah Alam, Selangor, Malaysia.

Keywords:

HSV, YCbCr, Canny Edge Detector, Otsu Method Threshold, Sobel Edge Detector.

Abstract

Hand gesture recognition system brings a lot of benefits to people in various industries. It consists of three main modules, which are the detection module, segmentation module and feature extraction module. The selection of appropriate segmentation method depends on the application and system environments. This paper concentrates on the following points: 1) The input images are converted to the HSV and YCbCr model by using threshold values, which satisfy the skin color segmentation based on skin color region; 2) Different edges operator are then applied on both color model images, namely Canny with fixed threshold, Canny with Otsu threshold and Sobel with fixed threshold. While there is no conclusion, which color space is best fit for skin color detection, Canny edge detector using Otsu threshold proves the best in creating contour of hand image since it is able to detect true weak edges compared to the other types of edge segmentation technique.

Downloads

Published

2017-06-01

How to Cite

Mod Ma’asum, F. F., Sulaiman, S., & Saparon, A. (2017). Comparative Study of Segmentation Technique for Hand Gesture Recognition System. Journal of Telecommunication, Electronic and Computer Engineering (JTEC), 9(2-2), 7–10. Retrieved from https://jtec.utem.edu.my/jtec/article/view/2211