Hand Gesture Recognition for Blind Using Machine Learning Algorithms

Abstract

Gesture recognition has been extremely important subfield under the Pattern Recognition stream which is currently receiving a lot of importance because of its potential applications. Gesture is a way by which a person makes use of his body and limbs for communicating an idea or feeling. In day-to-day life, people routinely communicate with one another with the use of hand gestures. Because of this, gestures form a natural way of communicating information. With this in mind, there has been a rise in interest for making use of gestures in communicating with computers. Taking into account human-computer interaction, gestures act as a possible alternative to traditional interface devices such as the keyboard, mouse, and joysticks (HCI). Taking all this into account, the main goal of gesture recognition is in developing a system that automatically recognizes and understands separate human movements and then utilize those gestures in communicating information. In this work, a user-friendly and effective smart hand gesture recognition is constructed at a cheap cost. As per the aspect ratio of the observed hand along its major - minor axis, palm and fist movements are identified with the help of machine learning algorithms .Making use of Viola-Jones procedure, the hand movement can be identified by analyzing both positive (hand) and negative (no hand) picture datasets.