bsuirДоклады БГУИРDoklady BGUIR1729-76482708-0382БГУИР10.35596/1729-7648-2023-21-2-114-120bsuir-3607Research ArticleЭЛЕКТРОНИКА, РАДИОФИЗИКА, РАДИОТЕХНИКА, ИНФОРМАТИКАELECTRONICS, RADIOPHYSICS, RADIOENGINEERING, INFORMATICSРаспознавание голоса с использованием свёрточной нейронной сетиVoice Detection Using Convolutional Neural NetworkВишняковВ. А.VishniakouU. A.

Минск

Vishniakou Uladzimir Anatolievich, Dr. of Sci. (Eng.), Professor at the Department of Infocommunication Technologies

220013, Minsk, P. Brovki St., 6

Tel.: +375 44 486-71-82

vish@bsuir.byШайяБ. Х.ShayaB. H.

Минск

Shaya Bahaa H., Postgraduate at the Department of Infocommunication Technologies

Minsk

Белорусский государственный университет информатики и радиоэлектроникиBelarusian State University of Informatics and Radioelectronics202324042023212114120Copyright © Вишняков В.А., Шайя Б.Х., 20232023Вишняков В.А., Шайя Б.Х.Vishniakou U.A., Shaya B.H.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://doklady.bsuir.by/jour/article/view/3607

Представлены подход, методология, программная система, основанные на свёрточной нейронной сети, для распознавания голоса (кашля) в условиях зашумленности с использованием технологий машинного обучения. Разработана и оценена система распознавания кашля на основе машинного обучения, использования свёрточной нейронной сети и библиотек языка Python. Свёрточная нейронная сеть протестирована с помощью различных наборов данных и библиотек. В отличие от существующих современных работ в этой области предложенная система оценивалась с применением реального набора звуковых данных окружающей среды, а не только отфильтрованных или разделенных звуковых параметров голоса. Окончательная скомпилированная модель показала относительно высокую среднюю точность – 85,37 %. Предлагаемая система способна распознавать звук голоса в многолюдном общественном месте, и нет необходимости в фазе разделения звука для предварительной обработки, как в других системах. Несколько добровольцев записали звуки своего голоса с помощью смартфонов. Затем они протестировали свои голоса в общественных местах на предмет шума в дополнение к некоторым аудиофайлам, которые были загружены онлайн. Результаты показали среднюю точность распознавания – 85,37 %, минимальную – 78,8 % и рекордную – 91,9 %. 

The article presents an approach, methodology, the software system based on a machine learning technologies for convolutional neural network and its use for voice (cough) recognition. Tasks of article are receiving evaluating a voice detection system with deep learning, the use of a convolutional neural network and Python language for patients with cough. The convolutional neural network has been developed, trained and tested using various datasets and Python libraries. Unlike the existing modern works related to this area, proposed system was evaluated using a real set of environmental sound data, and not only on filtered or separated voice audio tracks. The final compiled model showed a relatively high average accuracy of 85.37 %. Thus, the system is able to detect the sound of a voice in a crowded public place, and there is no need for a sound separation phase for pre-processing, as other modern systems require. Several volunteers recorded their voice sounds using microphones of their smartphones, and it was guaranteed that they would test their voices in public places to make noise, in addition to some audio files that were uploaded online. The results showed an average recognition accuracy – of 85.37 %, a minimum accuracy – of 78.8 % and a record – of 91.9 %. 

распознавание голосасвёрточная нейронная сетьнабор данных на основе машинного обученияаудиофайлыvoice detectionconvolution neural networkmachine learning-based datasetaudio filesReferencesShakel N. V., Ablameyko M. S. (2020) Medical Worker and Patient: Interaction in the Context of E-Health. Minsk, Eco-Perspective Publ. (in Russian).Shakel N. V., Ablameyko M. S. (2020) Medical Worker and Patient: Interaction in the Context of E-Health. Minsk, Eco-Perspective Publ. (in Russian).Alqudaihi K. S., Aslam N., Khan I. U. [et al.] (2021) Cough Sound Detection and Diagnosis Using Artificial Intelligence Techniques: Challenges and Opportunities. IEEE Public Health Emergency Collection. 9, 102327–102344.Alqudaihi K. S., Aslam N., Khan I. U. [et al.] (2021) Cough Sound Detection and Diagnosis Using Artificial Intelligence Techniques: Challenges and Opportunities. IEEE Public Health Emergency Collection. 9, 102327–102344.Amoh J., Odame K. (2016) Deep Neural Networks for Identifying Cough Sounds. IEEE Transactions on Biomedical Circuits and Systems. 10 (5), 1003–1011.Amoh J., Odame K. (2016) Deep Neural Networks for Identifying Cough Sounds. IEEE Transactions on Biomedical Circuits and Systems. 10 (5), 1003–1011.Gong Y., Lai C.-I. J., Chung Y.-A., Glass J. (2021) SSAST: Self-Supervised Audio Spectrogram Transformer. Applied Science. 570–575.Gong Y., Lai C.-I. J., Chung Y.-A., Glass J. (2021) SSAST: Self-Supervised Audio Spectrogram Transformer. Applied Science. 570–575.Nanni L., Maguolo G., Brahnam S., Paci M. (2021) An Ensemble of Convolutional Neural Networks for Audio Classification. Applied Science. 57–76.Nanni L., Maguolo G., Brahnam S., Paci M. (2021) An Ensemble of Convolutional Neural Networks forAudio Classification. Applied Science. 57–76.Chowdhury A., Ross A. (2019) Fusing MFCC and LPC Features using 1D Triplet CNN for Speaker Recognition in Severely Degraded Audio Signals. IEEE Transactions on Information Forensics and Security. 15, 1616–1629.Chowdhury A., Ross A. (2019) Fusing MFCC and LPC Features using 1D Triplet CNN for Speaker Recognition in Severely Degraded Audio Signals. IEEE Transactions on Information Forensics and Security. 15, 1616–1629.Visniakou U. A., Shaya B. H. (2022) Implementation of the Internet of Things Network for Monitoring Audio Information on a Microprocessor and Controller. System Analysis and Application Informatics. (1), 39–44.Visniakou U. A., Shaya B. H. (2022) Implementation of the Internet of Things Network for Monitoring Audio Information on a Microprocessor and Controller. System Analysis and Application Informatics. (1), 39–44.

The authors declare that there are no conflicts of interest present.