CB-27 Speech Enhancement with A Novel Deep-learning-based Model

SCURS Disciplines

Computer Sciences

Document Type

Poster Presentation

Abstract

Speech enhancement is a crucial aspect of audio signal processing, aimed at improving speech quality and intelligibility in noisy environments. Deep learning-based speech enhancement models have shown strong performance in noise reduction tasks. In this study, we extend DeepFilterNet, a speech enhancement model, by integrating AI-driven techniques to further refine speech quality. Our goal is to enhance noise suppression, improve speech clarity, and boost overall performance compared to the baseline model. The significance of this research lies in its contribution to speech enhancement, a crucial aspect of audio signal processing that impacts applications in telecommunications, hearing aids, voice assistants, and human-computer interaction. By refining DeepFilterNet with AI-driven techniques, this study aims to improve noise suppression and speech clarity, making speech more intelligible in challenging environments.

Keywords

Speech enhancement, machine learning, deep neural network

Start Date

11-4-2025 9:30 AM

Location

University Readiness Center Greatroom

End Date

11-4-2025 11:30 AM

This document is currently not available here.

Share

COinS
 
Apr 11th, 9:30 AM Apr 11th, 11:30 AM

CB-27 Speech Enhancement with A Novel Deep-learning-based Model

University Readiness Center Greatroom

Speech enhancement is a crucial aspect of audio signal processing, aimed at improving speech quality and intelligibility in noisy environments. Deep learning-based speech enhancement models have shown strong performance in noise reduction tasks. In this study, we extend DeepFilterNet, a speech enhancement model, by integrating AI-driven techniques to further refine speech quality. Our goal is to enhance noise suppression, improve speech clarity, and boost overall performance compared to the baseline model. The significance of this research lies in its contribution to speech enhancement, a crucial aspect of audio signal processing that impacts applications in telecommunications, hearing aids, voice assistants, and human-computer interaction. By refining DeepFilterNet with AI-driven techniques, this study aims to improve noise suppression and speech clarity, making speech more intelligible in challenging environments.