Пожалуйста, используйте этот идентификатор, чтобы цитировать или ссылаться на этот документ:
https://elib.bsu.by/handle/123456789/280033
Полная запись метаданных
Поле DC | Значение | Язык |
---|---|---|
dc.contributor.author | Kovalev, V. A. | |
dc.date.accessioned | 2022-05-24T13:13:31Z | - |
dc.date.available | 2022-05-24T13:13:31Z | - |
dc.date.issued | 2022 | |
dc.identifier.citation | Компьютерные технологии и анализ данных (CTDA’2022) : материалы III Междунар. науч.-практ. конф., Минск, 21–22 апр. 2022 г. / Белорус. гос. ун-т ; редкол.: В. В. Скакун (отв. ред.) [и др.]. – Минск : РИВШ, 2022. – С. 4-6. | |
dc.identifier.isbn | 978-985-586-561-3 | |
dc.identifier.uri | https://elib.bsu.by/handle/123456789/280033 | - |
dc.description | Секция «Системы машинного и глубокого обучения» | |
dc.description.abstract | In this paper, we study the problem of detecting typical vascular abnormalities of lungs that visually manifest themselves as a prominent vascularity of the roots of lungs. The study is capitalizing on a large dataset consisting of chest X-ray images of 15,600 people acquired as a result of computer-supported telemedicine screening of the population of subjects aged 18 years and older. The image training set consisted of 13,400 chest images, including 6,700 cases of pathology and 6,700 cases of the norm. The test set was composed from 2,200 images including 1,100 images of each class. Detecting vascular abnormalities was done by way of X-ray image classification using recent methods based on Convolutional Neural Networks. As a result, it was found that the presence of pathological changes can be recognized with the accuracy around 94% | |
dc.language.iso | en | |
dc.publisher | Минск : РИВШ | |
dc.rights | info:eu-repo/semantics/openAccess | |
dc.subject | ЭБ БГУ::ЕСТЕСТВЕННЫЕ И ТОЧНЫЕ НАУКИ::Кибернетика | |
dc.title | Detecting vascular abnormalities in lungs based on routine X-ray screening images and deep learning methods | |
dc.type | conference paper | |
Располагается в коллекциях: | 2022. Компьютерные технологии и анализ данных (CTDA’2022) |
Все документы в Электронной библиотеке защищены авторским правом, все права сохранены.