Hand Gesture Recognition system

Title

Hand Gesture Recognition system

Swapnil D.Badgujar, Gourab Talukdar, Omkar Gondhalekar, Mrs. S.Y. Kulkarni
Department of Computer, Engineering, Pune University, India
4 Assistant Prof., Department of Computer, Engineering, Pune University, India

Introduction

This research was done to identify human gestures, system architecture concepts for implementing a gesture recognition system, and major issues involved in implementing gesture recognition system.

Proposed system and purpose of the system

primary goal of this research was to create a system which can identify specific human gestures and use them to convey information or for device controls.
To implementation of this system real time hand tracking and extraction algorithm and feature extractions are used.
To gesture recognition use the vision sensors like cameras to acquire images, which are analysed to recognize the gestures.
purpose of this system was to control the traffic signals and mouse using hand gestures without using sensors at lower cost and with ease.

Working

system is going to be developed which can capture a hand gesture performed by the user in front of web Cam, this captured image is then processed to identify the valid gesture through specific algorithm & execute the corresponding operation.

Implementation 

1.Hand tracking and hand shape extraction

Here, a real-time hand tracking method is developed. This method is robust and reliable in complex background. For tracking the moving hand and then for extracting the hand shape fast and accurately, the trade-off between the computation complexity and robustness need to be considered.

2.Feature extraction

To find the movement information, the input gesture is assumed to be non-stationary or moving. When objects move in the spatial-time space, an image sequence is generated, motion detector is able to track the moving objects by examining the local Gray-level changes. Let Fi(x,y) be the ith frame of the sequence and Di(x,y) be the difference image between the ith and the (i+1)th frame defined as:

Di(x,y) = Ti{|Fi(x,y)-Fi+1(x,y)|}

where Ti is a thresholding function, Fi(x,y) and Di(x,y) are all 160 X 120 images, and Di(x,y) is binary image defined as follows:

3.Skin color detection

 Skin can be easily detected by using the color information. First, we use the constraint, RGB. some sample colors from the hand region may be obtained. To find the skin regions, we compare the colors in the regions with the prestored sample colors. If they are similar, then the region must be skin region.

4.Edge detection

 Edge detection is applied to separate the arm region from the hand region. There are fewer edges on the arm region than on the palm region. A simple edge detection technique, Kirsch edge operator to obtain different direction edges is used, and then the absolute maximum value of each pixel is chosen to form the edge image of ith frame as Ei(x,y):

5.Combination of motion, skin color, and edge

 The hand gestures information consists of skin color,movement and edge feature. We use the logic ‘AND’ to combine these three types of information

where Di(x,y), Si(x,y) and Ei(x,y) indicate the movement, skin color and edge images.

6.Region identification.

A simple method for region identification is to label each region with a unique integer number which is called the labelling process. After labelling, the largest integer label indicates the number of regions in the image. After the labelling process, the small regions can be treated as noise and then be removed

7.Background subtraction

 A simple background subtraction technique is used to obtain the hand gesture shape.
BGi+1 = (1-w)BGi + wFi

conclusion

In this paper they have presented a method to recognize the unknown input gestures by using hand tracking and extraction method. they apply this system to recognize the single gesture. In the experiments, they assume stationary background so that their system will have smaller search region for tracking.
Using this model they have developed an application where it can control mouse with the finger using it on web cam. Also they have developed an application controlling traffic signals using hand gestures.