Gesture Recognition: An Interactive Tool in Multimedia

Introduction

Gesture recognition aids human beings to commune with the machine (HMI) and interact in a natural manner without the use of any mechanical devices. Gesture recognition has replaced all input devices like joysticks, mouse and keyboards which have been acting as input devices to the system. For example, if you want to point at a particular diagram in a web page it is enough if we point it with our finger in place of mouse being used to do the same work, so this could be very useful for interactive purpose. Hand gestures can be used for natural and intuitive human-computer interaction. On the other hand, computer vision algorithms are notoriously brittle and computation intensive, which make most current gesture recognition systems fragile and inefficient. Its applications are diverse, spanning a variety of markets. To achieve this goal, computers should be able to visually recognize hand gestures from video input. However, vision-based hand tracking and gesture recognition is an extremely challenging problem due to the complexity of hand gestures, which are rich in diversities due to high degrees of freedom involved by the human hand. Now, gesture recognition is becoming a commonplace technology, enabling humans and computer to interface (HCI) more easily in homes, automobiles, work etc.

Over the past few years, gesture recognition has made its debut in entertainment and gaming markets. Gesture recognition (GR) has been presented by the RGB-D gaming controller which used a stereoscopic 2D image sensor which brought this GR technology to all the electronic devices in one’s life. This has altered the way with which users interact with the system. Gestures are useful in identifying hand and body movement. In addition, they could be used as input command which helps in reading the face and speech recognition. This provision of enabling gestures to be given as input commands to the computer makes it very easy and comfortable for the physically-impaired. Hand, body gesture can have modified amplification factor introduced by a controller which integrated with multiple sensors like accelerometer, gyroscope, etc., were connected to a camera to enable the software in the controller can recognize various gestures. Each and every gesture is (needed to) trained (always) for a particular function. For instance, still pose of the palm might indicate stop. In addition to the technical challenges of implementing gesture recognition, there are also social challenges. Gestures must be simple, intuitive and universally acceptable.

The study of gestures and other nonverbal types of communication is known as kinesics. 3D tracking and control software powers gesture control TVs, 3D depth sensing digital signs, digital signage and interactive hand-tracking displays. It caused a revolution in interactive technologies. Natural interaction seems to be the maximum in case of gesture recognition where there is possibility of gaming and 3-D virtual world environment. The whole world is revolving around the gesture recognition which is reaching the pinnacle in terms of the various applications that could be done with the help of it. It interacts tremendously with the real world. Process involved in gesture recognition: Gesture recognition which is an interactive technology and considered a very important tool in multimedia involves sequential process in its identification and recognition. Initially the acquisition of image is done with the help of standard camera or depth camera. This is followed by background subtraction where the foreground is separated from the background as the essential part of an image is only the foreground and noise if any present in the background need not be processed for its elimination. The foreground is then subject to filters depending on the kind of noise of present in the image. The output that is obtained after the filtering process (image free from noise) is then fed to hardware if any used for the purpose. Next in sequence is the use of classifier. There are a number of classifiers that could be used for identification and recognition of gestures they are Support Vector Machine, Neural Network, K - nearest neighbor etc.