REAL TIME HAND TRACKING &GESTURE RECOGNITION FOR HCI
      
      Cristina Manresa, Javier Varona, Ramon Mas and Francisco J. Perales

INTRODUCTION

• in this paper present a real time algorithm to track &recognise hand gestures for video game.

• the main goal of these algorithm is to measure the hand  configuration in each time . 

the algorithm based on three main steps

•        hand segentation
•        hand tracking
•        gesture recognition

these algorirhm only work in real time for reduced size hand  and  the hands holds most of the image.

METHODOLOGY

hand segmentation

          it is possible to use shapes.but they vary greatly during the hand motion of hand.there fore         use skin colour  as the hand feature.

the hand must be localized in the image &segmented from the background before recognition.the method is based in a probabilistics model of the skin colour pixel distribution.next the selected pixels 

are transfored from the RGB space to the HSL space.

tracking 

for the tracking use a USB camera and from this can cause errors in the hand segmentation proces.

and also add a tracking algorithm tries to maintain and propagate the hand state over time.

hand segmentation

this project gestures alphabet can be four hand directions.for avoiding fast hand gestures changes that  were not intended .every change should be kept fixed for 5 frames.

fro the hand contour and the hand convex hull can calculate  a sequence of contour points between 

two consecutive convex hull vertices.

results

 this project can be a good performanve,and also they can not track a fast movements.but cant properly recognise a movement  hand of right and front.

i think from this research  apply our project to hand track.