Automated Voice Quality Assessment System
Published:
Duration: TBD
Affiliation: Viettel Big Data Analytics Center, Viettel Telecom ยท Hanoi, Vietnam
Note: Full write-up coming soon.
Overview
Built an automated system for assessing voice quality across multiple perceptual dimensions (intelligibility, naturalness, noise level) using deep learning. Targets both clinical voice disorder detection and call-center audio quality monitoring.
Planned Content
- Multi-dimensional quality scoring model architecture
- Dataset construction and annotation methodology
- Evaluation against MOS (Mean Opinion Score) baselines
- Deployment pipeline for real-time call monitoring
Technologies
Python PyTorch Signal Processing ASR Speech Quality Deep Learning