Voice pathology classification based on High-Speed Videoendoscopy

Panek, Daria; Skalski, Andrzej; Zieliński, Tomasz P.; Deliyski, Dimitar D.

doi:10.1109/embc.2015.7318467

Cited by 5 publications

(5 citation statements)

References 26 publications

(24 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Others have used classifications based on a combination of the vocal fold shape and vascular pattern [34]. It has also been attempted to correlate voice pathologies with endoscopic videos of the vocal fold [23]. Furthermore, high-speed videos are analyzed with wavelet-based phonovibrograms and it is shown that a distinction between malignant and precancerous vocal fold lesions is possible [35].…”

Section: Related Workmentioning

confidence: 99%

A dataset of laryngeal endoscopic images with comparative study on convolution neural network-based semantic segmentation

et al. 2019

View full text Add to dashboard Cite

Purpose Automated segmentation of anatomical structures in medical image analysis is a prerequisite for autonomous diagnosis as well as various computer and robot aided interventions. Recent methods based on deep convolutional neural networks (CNN) have outperformed former heuristic methods. However, those methods were primarily evaluated on rigid, real-world environments. In this study, existing segmentation methods were evaluated for their use on a new dataset of transoral endoscopic exploration. Methods Four machine learning based methods SegNet, UNet, ENet and ErfNet were trained with supervision on a novel 7-class dataset of the human larynx. The dataset contains 536 manually segmented images from two patients during laser incisions. The Intersection-over-Union (IoU) evaluation metric was used to measure the accuracy of each method. Data augmentation and network ensembling were employed to increase segmentation accuracy. Stochastic inference was used to show uncertainties of the individual models. Patient-to-patient transfer was investigated using patient-specific fine-tuning. Results In this study, a weighted average ensemble network of UNet and ErfNet was best suited for the segmentation of laryngeal soft tissue with a mean IoU of 84.7 %. The highest efficiency was achieved by ENet with a mean inference time of 9.22 ms per image. It is shown that 10 additional images from a new patient are sufficient for patient-specific fine-tuning. Conclusion CNN-based methods for semantic segmentation are applicable to endoscopic images of laryngeal soft tissue. The segmentation can be used for active constraints or to monitor morphological changes and autonomously detect pathologies. Further improvements could be achieved by using a larger dataset or training the models in a self-supervised manner on additional unlabeled data.

show abstract

Section: Related Workmentioning

confidence: 99%

A dataset of laryngeal endoscopic images with comparative study on convolution neural network-based semantic segmentation

et al. 2019

View full text Add to dashboard Cite

show abstract

“…The left and right sides were determined automatically by fitting an ellipse on the segmented glottal area and assigning its major axis length as the midline of the glottal area (Panek et al, 2015). As can be observed in Figure 6, the glottal area technique is inappropriate for identifying paralysis of the vocal cords.…”

Section: Resultsmentioning

confidence: 99%

“…A number of approaches have been developed to quantify vocal cord motion such as Glottal Area Waveform or GAW (Panek et al, 2015;Woo, 2014;Gonzalez et al, 2013), phonovibrography (Lohscheller et al, 2008), kymography (Švec & Schutte, 2012), glottography (Karakozoglou et al, 2012), spatiotemporal analysis (Zhang et al, 2007) etc. Most of the research studies have focussed on quantitative assessment of vocal cord vibration during voice production.…”

Section: Introductionmentioning

confidence: 99%

Automatic Quantification of Vocal Cord Paralysis - An Application of Fibre-optic Endoscopy Video Processing

Menon

Petropoulakis

Soraghan

et al. 2017

Proceedings of the 10th International Joint Conference on Biomedical Engineering Systems and Technologies

View full text Add to dashboard Cite

Abstract:Full movement of the vocal cords is necessary for life sustaining functions. To enable correct diagnosis of reduced vocal cord motion and thereby potentially enhance treatment outcomes, it is proposed to objectively determine the degree of vocal cord paralysis in contrast to the current clinical practice of subjective evaluation. Our study shows that quantitative assessment can be achieved using optical flow based motion estimation of the opening and closing movements of the vocal cords. The novelty of the proposed method lies in the automatic processing of fibre-optic endoscopy videos to derive an objective measure for the degree of paralysis, without the need for high-end data acquisition systems such as high speed cameras or stroboscopy. Initial studies with three video samples yield promising results and encourage further investigation of vocal cord paralysis using this technique.

show abstract

“…However, recent work addressing these limitation using methods of machine learning, feature extraction, and automated analysis of HSV recordings have recently been shown to distinguish laryngeal pathologies with great accuracy. [35][36][37] Thus, it is likely that HSV imaging with automated multifeature objective analysis will become the standard method by which clinical laryngeal imaging and analysis will be conducted in the evaluation of voice disorders during the upcoming decade.…”

Section: High-speed Video Imagingmentioning

confidence: 99%

The Next 10 Years in Voice Evaluation and Treatment

Barkmeier-Kraemer

Patel

2016

Semin Speech Lang

View full text Add to dashboard Cite

Voice disorders are thought to affect approximately one third of all individuals within the United States during their lifetime. Individuals who require the use of their voice as part of their occupations are at highest risk for developing voice problems. Unfortunately, efficient diagnosis and effective management of voice disorders can be challenged by difficulty accessing professionals with the necessary expertise to diagnose and treat voice problems efficiently. Within the next decade, technological advancements show promise for improving the efficiency and effectiveness of intervention for voice disorders. Exciting developments in laryngeal imaging, modeling of patientspecific vocal patterns, and implementation of smart mobile technology and telehealth will greatly improve the accuracy of diagnosing voice problems and enhance implementation and carryover of effective voice treatment methods to daily communication demands. KEYWORDS: Voice, voice disorders, evaluation, treatment, technologyLearning Outcomes: As a result of this activity, the reader will be able to (1) identify and discuss current directions in laryngeal imaging technology; (2) identify and discuss current directions in developing patientspecific methods for diagnosis and treatment; and (3) identify and discuss future methods for implementing mobile smart technology and telehealth technology to address current challenges in voice treatment.Voice disorders affect nearly one-third of individuals in the United States age 20 to 66 years of age during their lifetime, with 39% of older people aged 45 to 64 years of age reporting voice problems.1 Of those diagnosed with voice disorders, two thirds are female.2 Twenty-eight million people in the United States are estimated to require their

show abstract

Voice pathology classification based on High-Speed Videoendoscopy

Cited by 5 publications

References 26 publications

A dataset of laryngeal endoscopic images with comparative study on convolution neural network-based semantic segmentation

A dataset of laryngeal endoscopic images with comparative study on convolution neural network-based semantic segmentation

Automatic Quantification of Vocal Cord Paralysis - An Application of Fibre-optic Endoscopy Video Processing

The Next 10 Years in Voice Evaluation and Treatment

Contact Info

Product

Resources

About