Webcam Based Hand Gesture Recognition System For Static Image

With the fast advancements of the internet and computer technologies since the1990s, the need for the application and the interest in doing research on Human computer interaction (HCI) has showed many improvements. It has progressed from conventional keyboard input to incorporation of the mouse and touch pad and latter to touch screen.Now, it is time to develop a much more convenient way to interact with computers. There are many sectors developing and applying seamless communication ways with computer systems. Either for gaming or for daily personal life use, the interaction of human with computers is done by the application of a sensor or a camera which will track the movement of human hand gestures, eye movement, facial expression or any other body parts movement. Of these other options hand gesture could be much more natural and intuitive way of communication with computer systems.In the past computers used to be considered as blinds. But now computers can see. They have an eye to see the outside environment as we do. The thing is, they need an advanced image processing methodology to be loaded on to them to further understand what they have seen.Noise elimination and lighting effects are the main problems for recognizing an object in image processing. So, to tackle it we have implemented background subtraction technique called code book, and other image processing algorithms like convex hull, convexity defect are implemented.This system will use any low end webcam to input the frames containing a hand gesture live from the person making the gesture. it will track the finger and palm position to recognize the hand gestures from continuous frames captured from a webcam, transform the input RGB color image into other color space, namely YCbCr and further process it to recognize the hand gesture from the final binary image which contains only two color values (white and black). With white showing the hand part and the black color representing the background area. This is by applying erosion on the frame to eliminate the noises and smoothing is by using dilating technique. Different scenarios have also been checked to make sure the hand recognition operates properly. The system demonstrated to function properly when the hand and arm are tilt, when the palm and arm are rotated, when fingers are folded. The hand gesture recognition interface proposed in this system can be applicable for sign language translation, playing virtual games, as TV or music player remote controller system, Human computer and cell phone interactions and so on.