Inclusive Interfaces: Enhancing Voice Assistants for Diverse Speech Abilities

Title

Inclusive Interfaces: Enhancing Voice Assistants for Diverse Speech Abilities

Creator

Pinto, Valencia; Kulkarni, Manasa; Govind, Sangeetha

Description

The main goal of this research is to highlight the considerable challenges that people suffering from dysarthria have to face while using Automatic Speech Recognition (ASR) technologies and virtual assistants. Current voice-controlled technologies struggle to understand disordered speech, which reveals a significant gap in research about how to adapt these technologies to speech changes caused by health conditions. We propose to explore non-verbal voice signals as a quieter potential solution. The work uses a systematic literature review and comparative analysis across various scenarios to point out the weaknesses of ASR in terms of accuracy and security. The research proves that non-verbal input always provides greater performance in accuracy, reliability, and convenience. This analysis lays the groundwork for designing more accessible technologies. Designed for researchers and developers, this research is an avenue for future work where future efforts address the empirical evaluation of a specialized non-verbal system to support user autonomy. 2025 IEEE.

Source

2025 17th IEEE International Conference on Computational Intelligence and Communication Networks, CICN 2025;pp.2028-2033

Date

01-01-2025

Publisher

Institute of Electrical and Electronics Engineers Inc.

Subject

Accessibility; Automatic Speech Recognition(ASR); Dysarthria; Non-Verbal Interaction; Voice Assistants

Coverage

Pinto V., CHRIST University, Department of Computer Science, Karnataka, Bangalore, India; Kulkarni M., CHRIST University, Department of Computer Science, Karnataka, Bangalore, India; Govind S., CHRIST University, Department of Computer Science, Karnataka, Bangalore, India

Rights

Restricted Access; Hardcopy may be available in the library

Relation

ISBN: 979-833158733-8;

Format

online

Language

English

Type

Conference paper

Identifier

https://doi.org/10.1109/CICN67655.2025.11368248

https://www.scopus.com/pages/publications/105034683407?origin=resultslist