Сибирский федеральный университет

Год издания: 2020

Идентификатор DOI: 10.1007/978-3-030-63319-6_36

Ключевые слова: contour analysis, descriptor, fourier, moment, ziehl-neelsen

Аннотация: The diagnosis of tuberculosis using automated systems in a quick and inexpensive method is relevant now, especially because tuberculosis is considered one of the most important public health problems worldwide. One of the components of this process is the task of creating a feature dictionary for the subsequent classification of mycobacterium tuberculosis on digital images of sputum stained by the Ziehl-Neelsen method. This paper considers the main principles of central moments, normalized central moments and hu moments, and discrete Fourier transform in relation to the contour analysis of studied objects. To compare the effectiveness of the selected methods, the neuro-fuzzy model of the Takagi Sugeno Kang algorithm on system AnFis is used. The feature vector was formed based on color descriptors (based on RGB model), shape descriptors (eccentricity and compactness), and contour descriptors. As comparison criteria, we used indicators of the regression coefficient, standard error, accuracy, specificity and sensitivity. © 2020, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG.

Журнал: Advances in Intelligent Systems and Computing

Выпуск журнала: Vol. 1295

Номера страниц: 397-403

ISSN журнала: 00253159

• Chentsov S.V. (Siberian Federal University, 79, Svobodny Ave., Krasnoyrsk, 660041, Russian Federation)
• Shelomentseva I.G. (Siberian Federal University, 79, Svobodny Ave., Krasnoyrsk, 660041, Russian Federation, Krasnoyarsk State Medical University Named After Professor V.F. Voino-Yasenetsky, 1, Partizan Zheleznyak Ave., Krasnoyrsk, 660022, Russian Federation)
• Yakasova N.V. (Siberian Federal University, 79, Svobodny Ave., Krasnoyrsk, 660041, Russian Federation, Khakassia State University Named After N. F. Katanov, 90, Lenin Ave., Abakan, Republic of Khakassia 655017, Russian Federation)

• Scopus
• Ядро РИНЦ (eLIBRARY.RU)