Author Login IJIRT Partner Network (IPN) Login

Advancing voice health

  • Unique Paper ID: 164723
  • Volume: 10
  • Issue: 12
  • PageNo: 2612-2618
  • Abstract:
  • In our study, we propose an innovative method for early detection and intervention of vocal disorders. Our comprehensive dataset consists of voice samples from healthy individuals and those with voice pathologies. We consider acoustic features like fundamental frequency, jitter, shimmer, and Mel-frequency cepstral coefficients, which are analyzed using tree-based machine learning algorithms. Additionally, we extract modes from audio signals through Variational Mode Decompo- sition (VMD) and convert them into Mel spectrograms. These spectrograms are then processed by a Vision Transformer archi- tecture. With a focus on multi-class classification, we combine the outputs of the tree-based algorithms and Vision Transformer into an ensemble model to enhance predictive accuracy across all classes. The method yields good results, which achieves an overall accuracy of 93% along with strong performance on other metrics, demonstrating its potential for improving early detection techniques for voice disorders.
Download article email to a friend

Cite This Article

  • ISSN: 2349-6002
  • Volume: 10
  • Issue: 12
  • PageNo: 2612-2618

Advancing voice health

Related Articles

Impact Factor
8.01 (Year 2024)
UGC Approved
Journal no 47859

Join Our IPN

IJIRT Partner Network

Submit your research paper and those of your network (friends, colleagues, or peers) through your IPN account, and receive 800 INR for each paper that gets published.

Join Now

Latest Publication

Archive

View More

Recent Conferences

NCSEM 2024

National Conference on Sustainable Engineering and Management - 2024 Last Date: 15th March 2024

Submit inquiry