تشخیص رویدادهای لکنت زبان در گفتار با استفاده از برچسب گذاری توالی
Abstract
1- Introduction
2- Previous work
3- Data transcription and annotation
4- Detecting stuttering events
5- Features of the classifiers used to detect stuttering events
6- Automatic speech recognition system
7- Experiments
8- Conclusion
References
Abstract
Stuttering is a speech disorder that, if treated during childhood, may be prevented from persisting into adolescence. A clinician must first determine the severity of stuttering, assessing a child during a conversational or reading task, recording each instance of disfluency, either in real time, or after transcribing the recorded session and analysing the transcript. The current study evaluates the ability of two machine learning approaches, namely conditional random fields (CRF) and bi-directional long-short-term memory (BLSTM), to detect stuttering events in transcriptions of stuttering speech. The two approaches are compared for their performance both on ideal hand-transcribed data and also on the output of automatic speech recognition (ASR). We also study the effect of data augmentation to improve performance. A corpus of 35 speakers’ read speech (13K words) was supplemented with a corpus of 63 speakers’ spontaneous speech (11K words) and an artificially-generated corpus (50K words). Experimental results show that, without feature engineering, BLSTM classifiers outperform CRF classifiers by 33.6%. However, adding features to support the CRF classifier yields performance improvements of 45% and 18% over the CRF baseline and BLSTM results, respectively. Moreover, adding more data to train the CRF and BLSTM classifiers consistently improves the results.
Introduction
Stuttering, also known as stammering, is a speech communication disorder that can have severe social, educational and emotional maladjustment consequences, not only for the people who stutter but also for their families [1, 2]. It is presumed that early intervention is best to offset potential later impacts of having a stutter on one’s psycho-social and communication developments [3]. During the assessment phase, clinicians carefully measure the stuttering events to determine if the stuttering is normal disfluency, borderline stuttering or beginning stuttering [4]. There are several approaches to determine stuttering severity. The fluency of very young children is commonly assessed through a conversational task, whereas for children older than seven years, a reading task may be used [5, 6]. The clinician asks the child to read from a passage, and then records each instance of disfluency while the child is reading. Clearly, this process is extremely dependent on the clinician’s experience [7, 8, 9, 10].