Analysis for Using Noise as a Source of Data Augmentation for Dysarthric Speech Recognition

Automatic speech recognition (ASR) for dysarthric speech is challenging. The acoustic characteristics of dysarthric speech are highly variable and there are often fewer distinguishing cues between phonetic tokens. Multimodal ASR utilises the data from other modalities to facilitate the task when a single acoustic modality proves insufficient. Articulatory information, which encapsulates knowledge about the speech production process, may constitute such a complementary modality. Although multimodal acoustic-articulatory ASR has received increasing attention recently, incorporating real articulatory data is under-explored for dysarthric speech recognition. This paper investigates the effectiveness of multimodal acoustic modelling using real dysarthric speech articulatory information in combination with acoustic features, especially raw signal representations which are more informative than classic features, leading to learning representations tailored to dysarthric ASR. In particular, various raw acoustic-articulatory multimodal dysarthric speech recognition systems are developed and compared with similar systems with hand-crafted features. Furthermore, the difference between dysarthric and typical speech in terms of articulatory information is systematically analysed by using a statistical space distribution indicator called Maximum Articulator Motion Range (MAMR). Additionally, we used mutual information analysis to investigate the robustness and phonetic information content of the articulatory features, offering insights that support feature selection and the ASR results. Experimental results on the widely used TORGO dysarthric speech dataset show that combining the articulatory and raw acoustic features at the empirically found optimal fusion level achieves a notable performance gain, leading to up to 7.6% and 12.8% relative word error rate (WER) reduction for dysarthric and typical speech, respectively. • Systematically compares the articulatory movement pattern mismatch between dysarthric and typical speech. This is achieved through the visualisation of the 3D point cloud of several real articulatory data samples and an analysis of the statistical space distribution across the dataset, employing the Maximum Articulator Motion Range (MAMR) as a metric. • Conduct comprehensive articulatory information analysis, including the computation of mutual information , to identify and select the most task-beneficial articulators, thereby enabling a more informed integration of articulatory features with acoustic features for optimal multimodal ADSR performance. • Investigate multimodal acoustic modelling for ADSR using real articulatory information in combination with acoustic features, encompassing both hand-crafted features and raw signal representations, e.g., raw waveform with non-parametric and parametric CNNs, magnitude spectrum, the raw source and filter components, and raw real and imaginary parts of the Fourier transform.

Acoustic modelling for automatic dysarthric speech recognition (ADSR) is a challenging task. Data deficiency is a major problem and substantial differences between typical and dysarthric speech complicate the transfer learning. In this paper, we aim at building acoustic models using the raw magnitude spectra of the source and filter components for ADSR. The proposed multi-stream models consist of convolutional, recurrent and fully-connected layers allowing for pre-processing various information streams and fusing them at an optimal level of abstraction. We demonstrate that such a multi-stream processing leverages information encoded in the vocal tract and excitation components and leads to normalising nuisance factors such as speaker attributes and speaking style. This leads to a better handling of dysarthric speech that exhibits large inter- and intra-speaker variabilities and results in a notable performance gain. Furthermore, we analyse the learned convolutional filters and visualise the outputs of different layers after dimensionality reduction to demonstrate how the speaker-related attributes are normalised along the pipeline. We also compare the proposed multi-stream model with various systems based on MFCC, FBank, raw waveform and i-vector, and, study the training dynamics as well as usefulness of the feature normalisation and data augmentation via speed perturbation. On the widely used TORGO and UASpeech dysarthric speech corpora, the proposed approach leads to a competitive performance of up to 35.3% and 30.3% WERs for dysarthric speech, respectively.

This research focuses on the automatic recognition and assessment of dysarthric speech, a critical area for individuals with speech impairments resulting from neurological conditions. The study explores advanced speech processing techniques, including deep learning models, autoencoder-based speech enhancement, and speaker adaptation approaches, to improve dysarthric speech recognition and intelligibility assessment. By leveraging both dysarthric and healthy speech corpora, the research enhances automatic speech recognition (ASR) accuracy, enabling more effective communication support for individuals with dysarthria. The societal impact of this work is significant. Automatic intelligibility assessment can assist speech therapists in tracking the progression of speech disorders and tailoring therapy plans, leading to improved rehabilitation outcomes. Enhanced ASR systems can also empower individuals with dysarthria by facilitating better communication in everyday interactions, reducing social isolation, and improving access to education and employment. This research contributes to the broader goal of inclusive technology, ensuring speech-impaired individuals can interact effectively with digital and human communication interfaces.

While dysarthric speech recognition can be a convenient interface for dysarthric speakers, it is hard to collect enough speech data to overcome the underestimation problem of acoustic models. In addition, there are lots of pronunciation variations in the collected database due to the paralysis of the articulator of dysarthric speakers. Thus, a discriminative training method is proposed for improving the performance of such resource-limited dysarthric speech recognition. The proposed method is applied to subspace Gaussian mixture modeling by incorporating pronunciation variations into a conventional minimum phone error discriminative training method.

Disabled individuals can realize many benefits from automatic speech recognition (ASR). To date, most ASR research has focused on normal speech. However, many individuals with physical disabilities also exhibit speech disorders. While limited research has been conducted focusing on dysarthric speech recognition, the preliminary results indicate that additional study is necessary. Recently, increasing attention has been given to multimodal speech recognition schemes that utilize multiple input sources-most commonly audio and video. This multimodal approach has been applied to normal speech with demonstrated effectiveness. Through studying the effect of audio and visual information in a human perception experiment, this study attempts to discover whether such an approach would be useful for dysarthric speech recognition. Results of a closed-vocabulary perception test are presented. In this test, 15 normal-hearing viewers were presented with video-tapes of three dysarthric speakers (with cerebral palsy) speaking a series of one-syllable nonsense words. These words differed only in the initial consonant. The words were presented in both audio-only and audio-visual modes. Perception rates in both modes were measured. The results are analyzed and compared to other studies of visual speech perception and dysarthric speech articulation.

Combined speech enhancement and auditory modelling for robust distributed speech recognition

In this paper, the improvements in the recently implemented Kannada speech recognition system is demonstrated in detail. The Kannada automatic speech recognition (ASR) system consists of ASR models which are created by using Kaldi, IVRS call flow and weather and agricultural commodity prices information databases. The task specific speech data used in the recently developed spoken dialogue system had high level of different background noises. The different types of noises present in collected speech data had an adverse effect on the on line and off line speech recognition performances. Therefore, to improve the speech recognition accuracy in Kannada ASR system, a noise reduction algorithm is developed which is a fusion of spectral subtraction with voice activity detection (SS-VAD) and minimum mean square error spectrum power estimator based on zero crossing (MMSE-SPZC) estimator. The noise elimination algorithm is added in the system before the feature extraction part. An alternative ASR models are created using subspace Gaussian mixture models (SGMM) and deep neural network (DNN) modeling techniques. The experimental results show that, the fusion of noise elimination technique and SGMM/DNN based modeling gives a better relative improvement of 7.68% accuracy compared to the recently developed GMM-HMM based ASR system. The least word error rate (WER) acoustic models could be used in spoken dialogue system. The developed spoken query system is tested from Karnataka farmers under uncontrolled environment.

An assistive system for persons with vocal impairment due to dysarthria converts dysarthric speech to normal speech or text. Because of the articulatory deficits, dysarthric speech recognition needs a robust learning technique. Representation learning is significant for complex tasks such as dysarthric speech recognition. We focus on robust representation for dysarthric speech recognition that involves recognizing sequential patterns of varying length utterances. We propose a hybrid framework that uses a generative learning based data representation with a discriminative learning based classifier. In this hybrid framework, we propose to use Example Specific Hidden Markov Models (ESHMMs) to obtain log-likelihood scores for a dysarthric speech utterance to form fixed dimensional score vector representation. This representation is used as an input to discriminative classifier such as support vector machine.The performance of the proposed approach is evaluatedusingUA-Speechdatabase.The recognitionaccuracy is much better than the conventional hidden Markov model based approach and Deep Neural Network-Hidden Markov Model (DNN-HMM). The efficiency of the discriminative nature of score vector representation is proved for "very low" intelligibility words.

E2E-DASR: End-to-end deep learning-based dysarthric automatic speech recognition

We have constructed a new speech data corpus consisting of the utterances of 221 elderly Japanese people (average age: 79.2) with the aim of improving the accuracy of automatic speech recognition (ASR) for the elderly. ASR is a beneficial modality for people with impaired vision or limited hand movement, including the elderly. However, speech recognition systems using standard recognition models, especially acoustic models, have been unable to achieve satisfactory performance for the elderly. Thus, creating more accurate acoustic models of the speech of elderly users is essential for improving speech recognition for the elderly. Using our new corpus, which includes the speech of elderly people living in three regions of Japan, we conducted speech recognition experiments using a variety of DNN-HNN acoustic models. As training data for our acoustic models, we examined whether a standard adult Japanese speech corpus (JNAS), an elderly speech corpus (S-JNAS) or a spontaneous speech corpus (CSJ) was most suitable, and whether or not adaptation to the dialect of each region improved recognition results. We adapted each of our three acoustic models to all of our speech data, and then re-adapt them using speech from each region. Without adaptation, the best recognition results were obtained when using the S-JNAS trained acoustic models (total corpus: 21.85% Word Error Rate). However, after adaptation of our acoustic models to our entire corpus, the CSJ trained models achieved the lowest WERs (entire corpus: 17.42%). Moreover, after readaptation to each regional dialect, the CSJ trained acoustic model with adaptation to regional speech data showed tendencies of improved recognition rates. We plan to collect more utterances from all over Japan, so that our corpus can be used as a key resource for elderly speech recognition in Japanese. We also hope to achieve further improvement in recognition performance for elderly speech.

Acoustic model adaptation using in-domain background models for dysarthric speech recognition

In this paper, we propose to use syllables as the acoustic units representing speech signals in an automatic speech recognition (ASR) system in order to improve the performance of the automatic recognition of dysarthric speech. The motivation behind using syllables comes from studies of human perception which demonstrate the central role of the syllable played in human perception and generation of speech. To test our proposed approach, a syllable-based speaker-independent HMM-based ASR system was designed using Hidden Markov Model Toolkit (HTK). A series of experiments on dysarthric speech has been carried out using a subset of NEMOURS database. The obtained results show that the relative improvement in the recognition rate using syllables were found to be 8.18% and 15.48% compared to the recognition rates obtained using monophones and triphones, respectively.

Dysarthric speech recognition is essential for enhancing communication and accessibility for individuals with speech impairments, yet its development is hindered by a scarcity of robust, speaker-specific datasets. This study explores low-resource dysarthric speech recognition through cross-speaker transfer using synthetic data and parameter-efficient fine-tuning (PEFT). We integrate SpeechT5 text-to-speech (TTS) synthesis with x-vector speaker embeddings to generate speaker-specific dysarthric speech, enabling model adaptation while preserving pathological speech characteristics such as prosodic irregularities. Experiments on the TORGO dataset show that mixed cross-synthetic data with LoRA fine-tuning achieves a WER of 0.17, representing a 71.7% improvement over the standard model (0.60 WER) without fine-tuning the TTS model. However, cross-dataset generalisation remains challenging, yielding higher WERs on MINDS-14 (4.69) and AMI (0.96–3.83) datasets. Whilst synthetic data enhances in-domain recognition, further research is needed to improve cross-dataset generalisation and speaker adaptation, particularly for low-resource pathological speech settings.

Pronunciation variation is a major problem in disordered speech recognition. This paper focus on handling the pronunciation variations in dysarthric speech by forming speaker-specific lexicons. A novel approach is proposed for identifying mispronunciations made by each dysarthric speaker, using state-specific vector (SSV) of phone-cluster adaptive training (Phone-CAT) acoustic model. SSV is low-dimensional vector estimated for each tied-state where each element in a vector denotes the weight of a particular monophone. The SSV indicates the pronounced phone using its dominant weight. This property of SSV is exploited in adapting the pronunciation of a particular dysarthric speaker using speaker-specific lexicons. Experimental validation on Nemours database showed an average relative improvement of 9% across all the speakers compared to the system built with canonical lexicon. Index Terms: Dysarthric speech recognition, phone-CAT, lexical modeling, pronunciations, phone confusion matrix

Children’s Speech Recognition (CSR) is a challenging task due to the high variability in children’s speech patterns and limited amount of available annotated children’s speech data. We aim to improve CSR in the often-occurring scenario that no children’s speech data is available for training the Automatic Speech Recognition (ASR) systems. Traditionally, Vocal Tract Length Normalization (VTLN) has been widely used in hybrid ASR systems to address acoustic mismatch and variability in children’s speech when training models on adults’ speech. Meanwhile, End-to-End (E2E) systems often use data augmentation methods to create child-like speech from adults’ speech. For adult speech-trained ASRs, we investigate the effectiveness of augmentation methods; speed perturbations and spectral augmentation, along with VTLN, in an E2E framework for the CSR task, comparing these across Dutch, German, and Mandarin. We applied VTLN at different stages (training/test) of the ASR and conducted age and gender analyses. Our experiments showed highly similar patterns across the languages: Speed Perturbations and Spectral Augmentation yield significant performance improvements, while VTLN provided further improvements while maintaining recognition performance on adults’ speech (depending on when it is applied). Additionally, VTLN showed performance improvement for both male and female speakers and was particularly effective for younger children.