Current Trends in Mass Communication(CTMC)

University of Jos, Department of Computer Science, Nigeria

Publications

• Research Article   
  An Intelligent Support System for Speech and Hearing-Impaired Persons Using Deep Learning for Nigerian Sign Language
  Author(s): Okorie Emmanuel O*, Nachamada V Blamah and Gideon Dadik Bibu
  
  This paper proposes an intelligent support system for speech- and hearing-impaired individuals using deep learning for Nigerian Sign Language (NSL). The system enables real-time communication between hearing/speech-impaired and normal individuals by converting spoken words into text for the impaired, and signs into audio for the normal person. The system utilizes a TensorFlow model trained using transfer learning to recognize NSL signs and convert them into audio signals. The model was trained using a dataset of English words converted to NSL, achieving an average confidence rate of 72.63%. The system comprises two main parts: a sign language recognition module and a speech recognition module. The sign language recognition module recognizes signs made by the hearing/speech-impaired person, while the speech recognition module recognizes words spoken by a normal person. The system has t.. Read More»
  

  Abstract HTML PDF