C Janott1, M Schmitt2, C Heiser3, W Hohenhorst4, M Herzog5, M Carrasco Llatas6, W Hemmert7, B Schuller2,8. 1. Munich School of BioEngineering, Technische Universität München, Boltzmannstraße 11, 85748, Garching, Deutschland. c.janott@gmx.net. 2. ZD.B Lehrstuhl für Embedded Intelligence for Health Care and Wellbeing, Universität Augsburg, Augsburg, Deutschland. 3. Hals-Nasen-Ohrenklinik und Poliklinik, Klinikum rechts der Isar, Technische Universität München, München, Deutschland. 4. Klinik für HNO-Heilkunde, Kopf- und Hals-Chirurgie, Alfried Krupp Krankenhaus, Essen, Deutschland. 5. Klinik für HNO-Krankheiten, Kopf- und Halschirurgie, Carl-Thiem-Klinikum Cottbus, Cottbus, Deutschland. 6. Servicio de Otorrinolaringología, Hospital Universitario Doctor Peset, Valencia, Spanien. 7. Munich School of BioEngineering, Technische Universität München, Boltzmannstraße 11, 85748, Garching, Deutschland. 8. Department of Computing, Imperial College London, London, Großbritannien.
Abstract
BACKGROUND: Acoustic snoring sound analysis is a noninvasive method for diagnosis of the mechanical mechanisms causing snoring that can be performed during natural sleep. The objective of this work is development and evaluation of classification schemes for snoring sounds that can provide meaningful diagnostic support. MATERIALS AND METHODS: Based on two annotated snoring noise databases with different classifications (s-VOTE with four classes versus ACLTE with five classes), identically structured machine classification systems were trained. The feature extractor openSMILE was used in combination with a linear support vector machine for classification. RESULTS: With an unweighted average recall (UAR) of 55.4% for the s‑VOTE model and 49.1% for the ACLTE, the results are at a similar level. In both models, the best differentiation is achieved for epiglottic snoring, while velar and oropharyngeal snoring are more often confused. CONCLUSION: Automated acoustic methods can help diagnose sleep-disordered breathing. A reason for the restricted recognition performance is the limited size of the training datasets.
BACKGROUND: Acoustic snoring sound analysis is a noninvasive method for diagnosis of the mechanical mechanisms causing snoring that can be performed during natural sleep. The objective of this work is development and evaluation of classification schemes for snoring sounds that can provide meaningful diagnostic support. MATERIALS AND METHODS: Based on two annotated snoring noise databases with different classifications (s-VOTE with four classes versus ACLTE with five classes), identically structured machine classification systems were trained. The feature extractor openSMILE was used in combination with a linear support vector machine for classification. RESULTS: With an unweighted average recall (UAR) of 55.4% for the s‑VOTE model and 49.1% for the ACLTE, the results are at a similar level. In both models, the best differentiation is achieved for epiglottic snoring, while velar and oropharyngeal snoring are more often confused. CONCLUSION: Automated acoustic methods can help diagnose sleep-disordered breathing. A reason for the restricted recognition performance is the limited size of the training datasets.
Entities:
Keywords:
Data analysis; Drug induced sleep endoscopy; Intrinsic sleep disorders; Obstructive sleep apnea; Respiratory signs and symptoms
Authors: Kun Qian; Christoph Janott; Vedhas Pandit; Zixing Zhang; Clemens Heiser; Winfried Hohenhorst; Michael Herzog; Werner Hemmert; Bjorn Schuller Journal: IEEE Trans Biomed Eng Date: 2016-10-21 Impact factor: 4.538
Authors: Christoph Janott; Clemens Heiser; Winfried Hohenhorst; Michael Herzog; Nicholas Cummins; Bjorn Schuller Journal: Conf Proc IEEE Eng Med Biol Soc Date: 2017-07
Authors: Christoph Janott; Maximilian Schmitt; Yue Zhang; Kun Qian; Vedhas Pandit; Zixing Zhang; Clemens Heiser; Winfried Hohenhorst; Michael Herzog; Werner Hemmert; Björn Schuller Journal: Comput Biol Med Date: 2018-01-31 Impact factor: 4.589
Authors: Clemens Heiser; Phillippe Fthenakis; Alexander Hapfelmeier; Sebastian Berger; Benedikt Hofauer; Winfried Hohenhorst; Eberhard F Kochs; Klaus J Wagner; Guenther M Edenharter Journal: Sleep Breath Date: 2017-03-31 Impact factor: 2.816