Touchless Written English Characters Recognition using Neural Network

Abstract

Touchless written English character recognizer (TER), a new touchless approach to write and an intelligent approach to recognize English characters has been proposed in this paper. In TER, the inputs of English characters have been taken by touchless fashion i.e. by sensing specific color object with a moving hand tracking in front of a webcam. Then they have been recognized by efficient Artificial Neural Network (ANN). Like the application of other traditional computer input devices such as mouse or keyboard, TER can be extended to write and recognize English words and sentences by adding characters one by one to the text editor. Proposed TER has been applied for several different forms of touchless writings, namely 26 English characters and 10 English digits. Here for training, ANN with Scale Conjugate Gradient (SCG) method has been used that converges the training time faster and recognizes with good generalization ability. TER can be useful for the disabled persons.