PURPOSE: The purpose of this study was to quantify sentence-level articulatory kinematics in individuals treated for oral squamous cell carcinoma (ITOC) compared to control speakers while also assessing the effect of treatment site (jaw vs. tongue). Furthermore, this study aimed to assess the relation between articulatory-kinematic measures and self-reported speech problems. METHOD: Articulatory-kinematic data from the tongue tip, tongue back, and jaw were collected using electromagnetic articulography in nine Dutch ITOC and eight control speakers. To quantify articulatory kinematics, the two-dimensional articulatory working space (AWS; in mm2), one-dimensional anteroposterior range of motion (AP-ROM; in mm), and superior-inferior range of motion (SI-ROM in mm) were calculated and examined. Self-reported speech problems were assessed with the Speech Handicap Index (SHI). RESULTS: Compared to a sex-matched control group, ITOC showed significantly smaller AWS, AP-ROM, and SI-ROM for both the tongue tip and tongue back sensor, but no significant differences were observed for the jaw sensor. This pattern was found for both individuals treated for tongue and jaw tumors. Moderate nonsignificant correlations were found between the SHI and the AWS of the tongue back and jaw sensors. CONCLUSIONS: Despite large individual variation, ITOC showed reduced one- and two-dimensional tongue, but not jaw, movements compared to control speakers and treatment for tongue and jaw tumors resulted in smaller tongue movements. A larger sample size is needed to establish a more generalizable connection between the AWS and the SHI. Further research should explore how these kinematic changes in ITOC are related to acoustic and perceptual measures of speech.

@en

Practice and recent evidence show that state-of-the-art (SotA) automatic speech recognition (ASR) systems do not perform equally well for all speaker groups. Many factors can cause this bias against different speaker groups. This paper, for the first time, systematically quantifies and finds speech recognition bias against gender, age, regional accents and non-native accents, and investigates the origin of this bias by investigating bias cross-lingually (i.e., Dutch and Mandarin) and for two different SotA ASR architectures (a hybrid DNN-HMM and an attention based end-to-end (E2E) model) through a phoneme error analysis. The results show that only a fraction of the bias can be explained by pronunciation differences between speaker groups, and that in order to mitigate bias, language- and architecture specific solutions need to be found.@en

In this paper, we build and compare multiple speech systems for the automatic evaluation of the severity of a speech impairment due to oral cancer, based on spontaneous speech. To be able to build and evaluate such systems, we collected a new spontaneous oral cancer speech corpus from YouTube consisting of 124 utterances rated by 100 non-expert listeners and one trained speech-language pathologist, which we made publicly available. We evaluated the systems in two scenarios: a scenario where transcriptions were available (reference-based) and a scenario where transcriptions might not be available (reference-free). The results of extensive experiments showed that (1) when transcriptions were available, the highest correlation with the human severity ratings was obtained using an automatic speech recognition (ASR) retrained with oral cancer speech. (2) When transcriptions were not available, the best results were achieved by a LASSO model using modulation spectrum features. (3) We found that naive listeners’ ratings are highly similar to the speech pathologist's ratings for speech severity evaluation. (4) The use of binary labels led to lower correlations of the automatic methods with the human ratings than using severity scores.

@en

Introduction: Surgical treatment for oral cancer leads to lasting changes of the vocal tract and individuals treated for oral cancer (ITOC) often experience speech problems. The purpose of this study was to analyse the acoustic properties of the spontaneous speech of individuals who were surgically treated for oral cancer. It was investigated (1) how key spectral measures of articulation change post-treatment; (2) whether changes are more related to target manner or place of articulation; and (3) how spectral measures develop at various time points following treatment. Method: A corpus consisting of 32.850 tokens was constructed by manually segmenting the speech of five (four female - one male) American English speaking ITOC. General acoustic characteristics (duration and spectral tilt), plosives (burst frequency), fricatives (centre of gravity and spectral skewness), and vowels (F1 and F2) were analysed using linear mixed effects regression and compared to control speech. Moreover, a within speaker analysis was performed for speakers with multiple recordings. Results: Manner of articulation is more predictive of post-treatment changes than place of articulation. Compared to controls, ITOC produced the fricatives /f, v, θ, ð, s, z, ʃ, ʒ/ with a lower centre of gravity while no differences were found for plosives and vowels. Longitudinal analyses show high within-speaker variation, but general improvements one-year post-treatment. Conclusions: Surgical oral cancer treatment changes the spectral properties of speech. Fricatives with varying manner of articulations were distorted, suggesting that manner of articulation is more predictive than place of articulation in identifying general problem areas for ITOC.

@en

In this paper, we introduce a new corpus of oral cancer speech and present our study on the automatic recognition and analysis of oral cancer speech. A two-hour English oral cancer speech dataset is collected from YouTube. Formulated as a low-resource oral cancer ASR task, we investigate three acoustic modelling approaches that previously have worked well with low-resource scenarios using two different architectures; a hybrid architecture and a transformer-based end-to-end (E2E) model: (1) a retraining approach; (2) a speaker adaptation approach; and (3) a disentangled representation learning approach (only using the hybrid architecture). The approaches achieve a (1) 4.7% (hybrid) and 7.5% (E2E); (2) 7.7%; and (3) 2.0% absolute word error rate reduction, respectively, compared to a baseline system which is not trained on oral cancer speech. A detailed analysis of the speech recognition results shows that (1) plosives and certain vowels are the most difficult sounds to recognise in oral cancer speech — this problem is successfully alleviated by our proposed approaches; (3) however these sounds are also relatively poorly recognised in the case of healthy speech with the exception of/p/. (2) recognition performance of certain phonemes is strongly data-dependent; (4) In terms of the manner of articulation, E2E performs better with the exception of vowels — however, vowels have a large contribution to overall performance. As for the place of articulation, vowels, labiodentals, dentals and glottals are better captured by hybrid models, E2E is better on bilabial, alveolar, postalveolar, palatal and velar information. (5) Finally, our analysis provides some guidelines for selecting words that can be used as voice commands for ASR systems for oral cancer speakers.

@en

Automatic speech recognition (ASR) systems have seen substantial improvements in the past decade; however, not for all speaker groups. Recent research shows that bias exists against different types of speech, including non-native accents, in state-of-the-art (SOTA) ASR systems. To attain inclusive speech recognition, i.e., ASR for everyone irrespective of how one speaks or the accent one has, bias mitigation is necessary. Here we focus on bias mitigation against non-native accents using two different approaches: data augmentation and by using more effective training methods. We used an autoencoder-based cross-lingual voice conversion (VC) model to increase the amount of non-native accented speech training data in addition to data augmentation through speed perturbation. Moreover, we investigate two training methods, i.e., fine-tuning and domain adversarial training (DAT), to see whether they can use the limited non-native accented speech data more effectively than a standard training approach. Experimental results show that VC-based data augmentation successfully mitigates the bias against non-native accents for the SOTA end-to-end (E2E) Dutch ASR system. Combining VC and speed perturbed data gave the lowest word error rate (WER) and the smallest bias against nonnative accents. Fine-tuning and DAT reduced the bias against non-native accents but at the cost of native performance.

@en

We present a voice conversion framework that converts normal speech into dysarthric speech while preserving the speaker identity. Such a framework is essential for (1) clinical decision making processes and alleviation of patient stress, (2) data augmentation for dysarthric speech recognition. This is an especially challenging task since the converted samples should capture the severity of dysarthric speech while being highly natural and possessing the speaker identity of the normal speaker. To this end, we adopted a two-stage framework, which consists of a sequence-to-sequence model and a nonparallel frame-wise model. Objective and subjective evaluations were conducted on the UASpeech dataset, and results showed that the method was able to yield reasonable naturalness and capture severity aspects of the pathological speech. On the other hand, the similarity to the normal source speaker’s voice was limited and requires further improvements.@en

In this paper, we investigate several existing and a new state-of-the-art generative adversarial network-based (GAN) voice conversion method for enhancing dysarthric speech for improved dysarthric speech recognition. We compare key components of existing methods as part of a rigorous ablation study to find the most effective solution to improve dysarthric speech recognition. We find that straightforward signal processing methods such as stationary noise removal and vocoder-based time stretching lead to dysarthric speech recognition results comparable to those obtained when using state-of-the-art GAN-based voice conversion methods as measured using a phoneme recognition task. Additionally, our proposed solution of a combination of MaskCycleGAN-VC and time stretching is able to improve the phoneme recognition results for certain dysarthric speakers compared to our time stretched baseline.

@en

In this paper, we propose a new approach to pathological speech synthesis. Instead of using healthy speech as a source, we customise an existing pathological speech sample to a new speaker’s voice characteristics. This approach alleviates the evaluation problem one normally has when converting typical speech to pathological speech, as in our approach, the voice conversion (VC) model does not need to be optimised for speech degradation but only for the speaker change. This change in the optimisation ensures that any degradation found in naturalness is due to the conversion process and not due to the model exaggerating characteristics of a speech pathology. To show a proof of concept of this method, we convert dysarthric speech using the UASpeech database and an autoencoder-based VC technique. Subjective evaluation results show reasonable naturalness for high intelligibility dysarthric speakers, though lower intelligibility seems to introduce a marginal degradation in naturalness scores for mid and low intelligibility speakers compared to ground truth. Conversion of speaker characteristics for low and high intelligibility speakers is successful, but not for mid. Whether the differences in the results for the different intelligibility levels is due to the intelligibility levels or due to the speakers needs to be further investigated.@en

Oral cancer speech is a disease which impacts more than half a million people worldwide every year. Analysis of oral cancer speech has so far focused on read speech. In this paper, we 1) present and 2) analyse a three-hour long spontaneous oral cancer speech dataset collected from YouTube. 3) We set baselines for an oral cancer speech detection task on this dataset. The analysis of these explainable machine learning baselines shows that sibilants and stop consonants are the most important indicators for spontaneous oral cancer speech detection.@en