DSSCNet: A Transfer Learning Framework for Cross-Corpus Dysarthric Speech Severity Classification

2026-06-20

AI Summary

This study addresses the challenges of speaker variability, class imbalance, and data scarcity in cross-corpus, speaker-independent classification of dysarthric speech severity. To tackle these issues, the authors propose DSSCNet, a deep neural network framework that integrates transfer learning with multi-corpus learning. The method enhances feature generalization and model robustness by pretraining on a source corpus and fine-tuning on the target corpus. Notably, this work pioneers the application of transfer learning to cross-corpus dysarthria assessment, achieving classification accuracies of 75.80% and 68.25% on the TORGO and UA-Speech datasets, respectively—substantially outperforming current state-of-the-art approaches and significantly reducing misclassification rates.

Abstract

Dysarthric speech severity classification is challenging due to speaker variability, class imbalance, and limited datasets. This study introduces DSSCNet, a deep learning model that employs transfer learning and multi-corpus learning to enhance speaker-independent classification. By pre-training on one dysarthric speech corpus and fine-tuning on another, DSSCNet achieves improved feature extraction and cross-corpus generalization. Experimental results demonstrate that DSSCNet outperforms state-of-the-art models for speaker-independent severity classification, achieving 75.80\% accuracy on TORGO and 68.25\% on UA-Speech, significantly reducing misclassification errors. The findings confirm that leveraging knowledge transfer between datasets improves model robustness, making DSSCNet well-suited for automated dysarthria assessment. This research contributes to the development of more effective assistive speech technologies for individuals with speech impairments.

Problem

• dysarthric speech
• severity classification
• cross-corpus
• speaker variability
• class imbalance

Innovation

• transfer learning
• cross-corpus learning
• dysarthric speech severity classification
• speaker-independent modeling
• deep neural networks