Object and Sign Detection System

Abstract: Communication can be defined as a act of exchanging information, Emotions, Feelings among each other or group of people. But in case of Dumb & Deaf people it becomes difficult to communicate. In this paper, a real time System for Sign Language detection was built through the images captured by PC camera. The main aim of this project is to help Disabled people, Dumb & Deaf, Paralyzed people to communicate with ease. This model detects the sign irrespective of the standard Sign Language. The existing Digital Models are slow, they take very plenty amount of time just to print a Alphabet, and thinking of whole sentence is a lot of time. This model overcomes the problem of time as it detects it as whole word other than a single alphabet. This model is proposed using TensorFlow Algorithm was made using a set of images for particular sign in different skin tones, lightning, and background, etc. The system displays high accuracy of 80-90% for the sign detection.

Keywords: Sign Language, Gestures, Real Time, Labeling Software, TensorFlow Object detection module.