Communicating effectively with the hearing impaired

• Main Author: Cheng, Eddy Kuan Quan
• Other Authors: Goh Wooi Boon
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; Natural language processing
• Online Access: https://hdl.handle.net/10356/180885
• Institution: Nanyang Technological University

This project investigates the potential of Natural Language Processing (NLP) models to allow effective communication for people with hearing difficulty, particularly in Singapore. This study focuses on speech summarisation as a tool to bridge the communication gaps, employing two widely used frameworks: SpaCy and NLTK models. The models are analysed to determine the critical components required for creating an effective speech summarisation tailored to the needs of people with hearing difficulty. Through systematic evaluation, the project identifies vital processes and evaluates the performance of the summaries in terms of comprehensibility using the Likert scale. The findings highlight the strengths and limitations of these models in producing clear, concise, and understandable summaries, emphasising accessibility for the hearing-impaired population. This research contributes to developing a speech summarisation model for the development of assistive technologies that can facilitate communication. The model is broken down and analysed to identify which parts of the NLP process are crucial in creating a speech summarisation model focusing on people with hearing difficulties in Singapore. A comprehensive evaluation was conducted using a Likert scale to assess the models’ effectiveness.