The College of Engineering Technology-Mosul, Northern Technical University discussed a master dissertation on an intelligent voice signature recognition system for family members to assist deaf children by the postgraduate student, Ms. Israa Mohammed Ibrahim Dawood.

The dissertation aimed at designing and implementing a low-cost, real-time intelligent system for recognizing the voices of a limited number of family members and this system utilizes signal processing and deep learning techniques, along with a multi-sensory alert mechanism that includes visual displays, light signals, and vibrations, tailored to the needs of deaf children within their home environment.

The dissertation reviewed that most available assistive technologies focus on sound amplification or environmental sound recognition without offering personalized solutions capable of directly and in real time distinguishing the voices of family members.

The dissertation employed a lightweight model combining fast Fourier transform and convolutional neural networks, implementing the system on a Raspberry Pi platform.

The dissertation demonstrated high accuracy in speaker recognition with a real-time response suitable for practical use, in addition to an emergency alert feature that allows the child to send a distress signal to their caregiver when needed. 

The dissertation recommended expanding the development of assistive systems for deaf children and the possibility of expanding the system in the future to include other environmental sounds, improving alert mechanisms, and integrating with modern communication technologies.