CC BY
11170
Секция 11
Список литературы
1. Q. Zhang, et al. Predisposing factors for recanalization of cerebral aneurysms after endovascular embolization: a multivariate study. J. of Neurolnterventional Surgery, 10(3):252-257, 2018.
An erythrocytes cell segmentation algorithm in medical images
A. A. Utyansky, D. S. Batishchev, E. S. Soynikova, V. M. Mikhelev Belgorod national research university Email: batishchev@bsu.edu.ru DOI: 10.24411/9999-017A-2020-10299
This paper deals with the problem of segmentation of red blood cells in the peripheral blood pictures. A histogram of their distribution (the Price-Jones curve) is constructed from the diameters of the segmented red blood cells, and the curve can indicate the presence of certain diseases, or simply focus on an abnormal image. The original images contain distortion, noise, as well as the combined cells. So, first we need to suppress noise, divide the group into individual objects, and plot a curve through them.
Pre-filtering is performed in several steps - Gaussian blur to suppress high-frequency noise, equalization of contrast with the CLAHE algorithm.
The described segmentation algorithm is built on the method of dividing points. It also consists of several steps after preprocessing: find the contours of all cells (both single and group), group contours, evaluating and extracting contours, grouping segments of contours and fitting ellipses.
This work was financially supported by the Russian Foundation for Basic Research (project code 19-07-00133).
Competition and collaboration in the miRNA science field
A. B. Firsov12,1.1. Titov,3 1 Institute of Informatics Systems 2Novosibirsk State University 3Institute of Cytology and Genetics SB RAS Email: artemijfirsov@mail.ru, titov@bionet.nsc.ru DOI: 10.24411/9999-017A-2020-10300
Many digital libraries, such as PubMed/Scopus provide us with the opportunity to query articles metadata. This includes estimating the authors [1] institutions activity, revealing their interactions and other properties. We present the analysis of the institution's interactions in the miRNA science field using the PubMed digital library data. To tackle the problem of the affiliation variability [2], we proposed the k-mer boolean feature vector sorting algorithm - KOFER.
We identified the leaders of the field, characterized the interactions and described the country level features of co-authorship. We also provide the approximation of the miRNA science field, showing that the fields' peak is yet to be reached. We compare the publications activity patterns on the organization level, and provide additional insights of miRNA science field evolution.
References
1. А. А. Блинов И. И. Титов. Исследование структуры и эволюции сетей научного соавторства на основе анализа новосибирских публикаций в области биологии и медицины // Вавиловский журнал генетики и селекции, 18.4/2, 2014.
2. Shu Zhang et al. "An Adaptive Method for Organization Name Disambiguation with Feature Reinforcing". PACLIC, Bali, Indonesia, 2012.
Выделение спектральных серий в тандемных масс спектрах с помощью машинного обучения
Э. С. Фомин, Н. А. Алемасов Институт цитологии и генетики СО РАН Email: fomin@bionet.nsc.ru DOI: 10.24411/9999-017A-2020-10301
Огромный поток данных генерируемый масс спектрометрией для решения задач протеомики требует адекватных автоматических инструментов их анализа. При этом большая (до 90 %) часть данных является шумом и/или избыточна для задач идентификации и секвенирования [1], где достаточно
