Enhancing Speaker-Independent Dysarthric Speech Severity Classification with DSSCNet and Cross-Corpus Adaptation
Abstract
Dysarthric speech severity classification is crucial for objective clinical assessment and progress monitoring in individuals with motor speech disorders. Although prior methods have addressed this task, achieving robust generalization in speaker-independent (SID) scenarios remains challenging. This work introduces DSSCNet, a novel deep neural architecture that combines Convolutional, Squeeze-Excitation (SE), and Residual network, helping it extract discriminative representations of dysarthric speech from mel spectrograms. The addition of SE block selectively focuses on the important features of the dysarthric speech, thereby minimizing loss and enhancing overall model performance. We also propose a cross-corpus fine-tuning framework for severity classification, adapted from detection-based transfer learning approaches. DSSCNet is evaluated on two benchmark dysarthric speech corpora: TORGO and UA-Speech under speaker-independent evaluation protocols: One-Speaker-Per-Severity (OSPS) and Leave-One-Speaker-Out (LOSO) protocols. DSSCNet achieves accuracies of 56.84% and 62.62% under OSPS and 63.47% and 64.18% under LOSO setting on TORGO and UA-Speech respectively outperforming existing state-of-the-art methods. Upon fine-tuning, the performance improves substantially, with DSSCNet achieving up to 75.80% accuracy on TORGO and 68.25% on UA-Speech in OSPS, and up to 77.76% and 79.44%, respectively, in LOSO. These results demonstrate the effectiveness and generalizability of DSSCNet for fine-grained severity classification across diverse dysarthric speech datasets.
The following analysis is generated by AI and may contain errors. It is separate from the authors’ abstract and is not peer review.
Research summary
The main contribution of this paper is the development of DSSCNet, a novel deep learning architecture that significantly improves dysarthric speech severity classification through innovative use of transfer learning and advanced neural network components. This work represents a meaningful step forward in the field of speech processing, addressing critical challenges in speaker-independent scenarios and enhancing the potential for real-world applications.
Methodology
The proposed DSSCNet architecture effectively integrates convolutional layers, squeeze-excitation blocks, and residual connections to enhance the classification of dysarthric speech severity. The methodology is well-structured, leveraging deep learning principles to address the challenges of speaker-independent classification. The incorporation of a cross-corpus fine-tuning strategy is particularly noteworthy, as it allows the model to generalize better across different datasets, which is a significant advancement over traditional methods that often struggle with speaker variability.
Results & evaluation
The experimental setup is robust, utilizing two benchmark datasets (TORGO and UA-Speech) and employing rigorous evaluation protocols (OSPS and LOSO) to assess model performance. The results demonstrate a clear improvement over baseline models and existing state-of-the-art methods, particularly after fine-tuning, which showcases the effectiveness of the proposed architecture. However, the paper could benefit from more extensive ablation studies to further validate the contributions of individual components.
Reproducibility
The paper provides a detailed description of the methodology, including data preprocessing, model architecture, and training procedures, which aids in reproducibility. However, the absence of a publicly available code repository or demo URL limits the ability for other researchers to replicate the results directly.
Limitations
While the model shows promising results, it still faces challenges related to class imbalance, particularly for medium severity levels. The reliance on two specific datasets may also limit the generalizability of the findings to broader dysarthric speech scenarios. Additionally, the paper does not address potential overfitting issues, which could arise from the model's complexity.
Broader impact
The implications of this research are significant for clinical applications, particularly in developing assistive technologies for individuals with dysarthria. By improving the accuracy of severity classification, the proposed model can enhance the effectiveness of therapy plans and assistive communication devices, ultimately contributing to better quality of life for affected individuals. The approach also lays the groundwork for future research in speaker-independent speech processing tasks. The main contribution of this paper is the development of DSSCNet, a novel deep learning architecture that significantly improves dysarthric speech severity classification through innovative use of transfer learning and advanced neural network components. This work represents a meaningful step forward in the field of speech processing, addressing critical challenges in speaker-independent scenarios and enhancing the potential for real-world applications.
Author of this paper? Send a correction.