Repetition Detection in Stuttered Speech

Ramteke, Pravin Bhaskar; Koolagudi, Shashidhar G.; Afroz, Fathima

doi:10.1007/978-81-322-2538-6_63

Cited by 12 publications

(6 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another obvious difficulty with spontaneous speech is that, contrary to read speech, no reference script is available. To quantify the number of repetitions in spontaneous speech (i.e., to calculate the additional predictor aforementioned), advanced signal-processing algorithms should therefore be used, as those developed for the automatic detection of syllable repetitions in stuttered speech (Chee et al, 2009;Ramteke et al, 2016;Sahidullah et al, 2023).…”

Section: Discussionmentioning

confidence: 99%

Automatically measuring speech fluency in people with aphasia: first achievements using read-speech data

Fontan¹,

Prince

Nowakowska

et al. 2023

Aphasiology

View full text Add to dashboard Cite

Background. Speech and language pathologists (SLPs) often rely on judgements of speech fluency for diagnosing or monitoring patients with aphasia. However, such subjective methods have been criticised for their lack of reliability and their clinical cost in terms of time.Aims. This study aims at assessing the relevance of a signal-processing algorithm, initially developed in the field of language acquisition, for the automatic measurement of speech fluency in people with aphasia (PWA).Methods & Procedures. Twenty-nine PWA and five control participants were recruited via non-profit organizations and SLP networks. All participants were recorded while reading out loud a set of sentences taken from the French version of the Boston Diagnostic Aphasia Examination. Three trained SLPs assessed the fluency of each sentence on a five-point qualitative scale. A forward-backward divergence segmentation and a clustering algorithm were used to compute, for each sentence, four automatic predictors of speech fluency: pseudosyllable rate, speech ratio, rate of silent breaks, and standard deviation of pseudo-syllable length. The four predictors were finally combined into multivariate regression models (a multiple linear regression -MLR, and two non-linear models) to predict the average SLP ratings of speech fluency, using a leave-one-speaker-out validation scheme.Outcomes & Results. All models achieved accurate predictions of speech fluency ratings, with average root-mean-square errors as low as 0.5. The MLR yielded a correlation coefficient of 0.87 with reference ratings at the sentence level, and of 0.93 when aggregating the data for each participant. The inclusion of an additional predictor sensitive to repetitions improved further the predictions with a correlation coefficient of 0.91 at the sentence level, and of 0.96 at the participant level. Conclusions.The algorithms used in this study can constitute a cost-effective and reliable tool for the assessment of the speech fluency of patients with aphasia in read-aloud tasks.Perspectives for the assessment of spontaneous speech are discussed.

show abstract

Section: Discussionmentioning

confidence: 99%

Automatically measuring speech fluency in people with aphasia: first achievements using read-speech data

Fontan¹,

Prince

Nowakowska

et al. 2023

Aphasiology

View full text Add to dashboard Cite

show abstract

“…In subsequent research, many scholars have also used ANNs as classifiers for identifying stuttering events in speech [41,42]. In recent years, many researchers have used an increasing number of machine learning models as classifiers for detecting disfluency events in speech, such as hidden Markov models (HMM) [16,26], support-vector machines (SVM) [17,21,27,43], k-nearest neighbors (KNN) [18,20,22,27,28], linear discriminant analysis (LDA) [18,20,27,28], dynamic time warping (DTW) [19,44], and multilayer perceptrons (MLP) [45,46]. When using machine learning models for stuttering event classification, it is generally necessary to manually design some features to represent different aspects of speech, such as the spectrum, energy, and speaking rate of the audio.…”

Section: Disfluency Detection With Machine Learningmentioning

confidence: 99%

Automatic Speech Disfluency Detection Using wav2vec2.0 for Different Languages with Variable Lengths

et al. 2023

View full text Add to dashboard Cite

Speech is critical for interpersonal communication, but not everyone has fluent communication skills. Speech disfluency, including stuttering and interruptions, affects not only emotional expression but also clarity of expression for people who stutter. Existing methods for detecting speech disfluency rely heavily on annotated data, which can be costly. Additionally, these methods have not considered the issue of variable-length disfluent speech, which limits the scalability of detection methods. To address these limitations, this paper proposes an automated method for detecting speech disfluency that can improve communication skills for individuals and assist therapists in tracking the progress of stuttering patients. The proposed method focuses on detecting four types of disfluency features using single-task detection and utilizes embeddings from the pre-trained wav2vec2.0 model, as well as convolutional neural network (CNN) and Transformer models for feature extraction. The model’s scalability is improved by considering the issue of variable-length disfluent speech and modifying the model based on the entropy invariance of attention mechanisms. The proposed automated method for detecting speech disfluency has the potential to assist individuals in overcoming speech disfluency, improve their communication skills, and aid therapists in tracking the progress of stuttering patients. Additionally, the model’s scalability across languages and lengths enhances its practical applicability. The experiments demonstrate that the model outperforms baseline models in both English and Chinese datasets, proving its universality and scalability in real-world applications.

show abstract

“…To enhance classification accuracy, a decision fusion www.ijacsa.thesai.org technique is introduced, based on combination of different acoustical features like ZCR, speech envelope (ENV) for classifying filled pause (FP) and elongation (ELO) in Malay language [7]. Stuttered speech repetition detection algorithm based on MFCC and dynamic time warping (DTW) with accuracy of 83-90% is proposed in [8], [9].…”

Section: Lp-hilbert Transform Based Mfcc (Lh-mfcc) Basedmentioning

confidence: 99%

“…4 shown above describes a set of triangular filters to enumerate the weighted sum of all filter spectral samples so that the output is made to approach the Mel scale. Every filter has triangular magnitude frequency response with unit value at the center frequency and it gradually reduces linearly to zero at the Centre frequency of adjoining filters [7], [8]. Output of each filter is the filtered sum of its spectral components.…”

Section: B Mel-frequency Cepstral Coefficients (Mfccs)mentioning

confidence: 99%

MFCC and Texture Descriptors based Stuttering Dysfluencies Classification using Extreme Learning Machine

Jegan¹,

Jayagowri²

2022

IJACSA

View full text Add to dashboard Cite

Stuttering is a type of speech disorder which results in disrupted flow of speech in the form of unintentional repetitions and prolongation of sounds. Stuttering classification is important for speech pathology treatment and speech therapy techniques which decreases speech disfluency to some extent. In this article, a method for prolongation and repetition classification is presented based on Mel-frequency cepstral coefficients (MFCC) and texture descriptors. Initially, MFCC and filter bank energy (FBE) matrix are computed. Gray level co-occurrence matrix (GLCM) and Gray level run length matrix (GLRLM) textural features are extracted from these matrices. Laplacian score-based feature selection approach is employed to choose relevant features. Finally, extreme learning machine (ELM) is utilized to classify the speech audio event as repetition or prolongation. The algorithm is evaluated using UCLASS database and has achieved improved performance with classification accuracy of 96.36%.

show abstract

Repetition Detection in Stuttered Speech

Cited by 12 publications

References 7 publications

Automatically measuring speech fluency in people with aphasia: first achievements using read-speech data

Automatically measuring speech fluency in people with aphasia: first achievements using read-speech data

Automatic Speech Disfluency Detection Using wav2vec2.0 for Different Languages with Variable Lengths

MFCC and Texture Descriptors based Stuttering Dysfluencies Classification using Extreme Learning Machine

Contact Info

Product

Resources

About