System Design and Implementation of Assistive Device for Hearing Impaired People

This paper discusses the design, implementation and development of assistive device for hearing impaired. Sign language is one of the oldest and most natural form of language for communication, but since most people does not know sign language. Finding interpreters for different group of people is difficult. To treat the deaf people as one among the society, a user-friendly assistive device is necessary. This paper provides the details about the implementation of standalone interpreter using transfer learning for finger spelling-based American Sign Language using Raspberry Pi. A Graphical User Interface (GUI) is created and tested for establishing a two-way communication to convert text into sign language. A language barrier is created using an assign language structure that is different from normal text. Hence, the users will now depend on the vision-based communication that will be able to bring normal people, deaf and mute people on the same grounds of interaction. There is a need to build a possible sign language translator, which can take communication in sign language and translate them into written and oral language. Such a translator that uses neural networks for finger spelling-based American Sign Language would greatly lower the barrier for many deaf and mute individuals to be able to better communicate with others in day-to-day interactions. In proposed method, hand gesture is first passed through a filter, and after the filter has applied, the gesture is passed through a classifier which predicts the class of the hand gestures. This method provides 95.7% accuracy for the 26 alphabets and 0, 1, 2 numerical.