Sign Language Translation System for Railway Station Announcements

Abstract

The research aims at developing a system for hearing impaired citizens at public platforms such as railway stations, banks and bus stands, where in information transfer is crucial. Information at these platforms are generally conveyed either through text displays or audio announcements and the main drawback of conventional information broadcast is that they are not perceived by deaf people due to their inability to read or hear. In this research work, a generic sign language converter has been implemented for railway announcements in India. The corpus generation algorithm that provides the data of root words has been implemented thoroughly to provide legitimate database. For the text translation, effective and efficient use of phrase-based technique combined with rule-based translation technique has been used that yields optimum speed output thereby reducing redundancies and thus time for translation. The animated avatar has Inverse Kinematic(IK) solver tools and other rigged joints and chains which aids in movement of body parts similar to human. The sentence is translated into ASL gloss and the Web UI provides an interface which is easy to operate, it displays video of an avatar which conveys information using hand signs.