Novel Method for using Hand Recognition as Computer Remote Control with Computer Vision Techniques

Abstract:

Today, interaction between man-computer (HCI) is one of the most prominent goals. One of the important goals is to develop an independent control of external devices or static controls over a computer for simplified system and userfriendly interface. Detection and recognition of gestural parts of a person's hand plays a crucial role because it is used to perform almost all of the daily activities. This work is aimed at facilitating the way of exercising control over the PC using C++ programming language via machine vision libraries like OpenCV. The segmentation of the hand was performed in two stages: the first stage used a range of color with the HSV model, accompanied by morphological operations to minimize noise. The second stage, which was conducted after binarization, continued to seek the contour of the hand focusing on the most important features of its geometry. A convex hull and convexity defects were set to determine the type of gesture and assign a particular function to run on the computer. The center of the mass of the contour of the hand was located to obtain its coordinates (x, y). It was subsequently assigned to the mouse position connected to the PC to emulate the 2D scrolling on the screen, consistent to the movement of the hand. In comparison to the traditional input devices, this approach facilitated a convenient manipulation of computer tools, providing a greater control and user comfort.

Año de publicación:

2017

Keywords:

Fuente:

google