Speech-Based Parkinson’s Detection Using Pre-Trained Self-Supervised Automatic Speech Recognition (ASR) Models and Supervised Contrastive Learning

Speech-Based Parkinson’s Detection Using Pre-Trained Self-Supervised Automatic Speech Recognition (ASR) Models and Supervised Contrastive Learning

Diagnosing Parkinson’s disease (PD) through speech analysis is a promising area of research, as speech impairments are often one of the early signs of the disease. This study investigates the efficacy of fine-tuning pre-trained Automatic Speech Recognition (ASR) models, specifically Wav2Vec 2.0 and...

Full description

Saved in:
Bibliographic Details
Main Authors: Hadi Sedigh Malekroodi, Nuwan Madusanka, Byeong-il Lee, Myunggi Yi
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Bioengineering
Subjects:
Parkinson’s disease (PD)
deep learning
transfer learning
supervised contrastive learning
Wav2Vec 2.0
HuBERT
Online Access:https://www.mdpi.com/2306-5354/12/7/728
Tags: Add Tag
No Tags, Be the first to tag this record!

Similar Items

Similar Items