Color Based Hand and Finger Detection Technology for User Interaction

Title

Color Based Hand and Finger Detection Technology for User Interaction

Sung Kwan Kang, Mi Young Nam , Phill Kyu Rhee

Dept. of Computer Science & Engineering, Inha University(2008)

Introduction

The aim of this paper is to present the methodology for hand detection and propose the finger detection method. Skin color is used to segment the hand region from background and counter is extracted from the segmented hand. Analysis of counter gives us the location of finger tip in the hand.

Methodology

In this section we explain our method divided in three main steps.

1.     .  Hand segmentation
2.        Couture extraction
3.        Finger recognition

Color segmentation

Color Segmentation is done by analyzing the skin color over the skin color range. The simple model of skin color indicates that red component of the skin color is in the range of 37 and 60, whereas the green component of the skin color is between 28 and 34.

The transformation is done by software conversion, we use YUV for less expensive conversion from RGB. We use the OpenCV implementation (OpenCV 2005) which takes advantage of MMX instruction to perform faster RGB to YUV conversion.

Contour Extraction

Contour Extraction is the process of retrieving the location of pixel value of the edges of the detected objects from an image.

Once the hand blob is detected then, the edge vector of the particular blob can be retrieved using Edges () function of Blob class.

Refined contours can be obtained by minimizing the snake energy of each of the returned contours with respect to the binary image.

Finally, we obtain the blob representation of the hand.

 Contour Analysis and Finger Detection

Once we have contour vector, i.e. the silhouette of the hand region, the vector can be analyzed to find the finger location.

In the current version of application, only one finger whose height is maximum among others is found by contour analysis. The contour vector contains the x and y coordinates of the each point of hand silhouette. First maximum value of y coordinate is searched and the corresponding x coordinate corresponding to y coordinate is used to locate the finger tip.

To find the fingers blob in an image, morphological opening can be used.

 Conclusion

A color based method for hand location is presented and hand contour analysis is done to locate the finger tip.

Color detection algorithm is simple and more robust method like probabilistic method of color modeling. Once the hand counter is extracted from hand blob, highest position of contour is considered as finger tip.