informИнформатикаInformatics1816-03012617-6963UIIP NASBinform-474Research ArticleОБРАБОТКА СИГНАЛОВ, ИЗОБРАЖЕНИЙ, РЕЧИ, ТЕКСТА И РАСПОЗНАВАНИЕ ОБРАЗОВSIGNAL, IMAGE, SPEECH, TEXT PROCESSING AND PATTERN RECOGNITIONОтбор информативных геометрических признаков ядер клеток на люминесцентных изображениях раковых клетокSelection of geometrical features of nuclei оn fluorescent images of cancer cellsЛисицаЕ. В.LisitsaYa. U.

Лисица Евгения Владимировна, научный сотрудник

факультет радиофизики и компьютерных технологий

Yauheniya U. Lisitsa, Researcher, the Faculty of Radiophysics and Computer Technologies

ylisitsa@gmail.comЯцковН. Н.YatskouM. M.

Яцков Николай Николаевич, кандидат физикоматематических наук, доцент

факультет радиофизики и компьютерных технологий

Mikalai M. Yatskou, Cand. Sci. (Phys.-Math.), Assoc. Prof., the Faculty of Radiophysics and Computer Technologies

yatskou@bsu.byСкакунВ. В.SkakunV. V.

Скакун Виктор Васильевич, кандидат физикоматематических наук, доцент

факультет радиофизики и компьютерных технологий

Victor V. Skakun, Cand. Sci. (Phys.-Math.), Assoc. Prof., the Faculty of Radiophysics and Computer Technologies

skakun@bsu.byКривошеевП. Д.Pavel D. KryvasheyeuP. D.

Кривошеев Павел Дмитриевич, студент

факультет радиофизики и компьютерных технологий

Pavel D. Kryvasheyeu, Student, the Faculty of Radiophysics and Computer Technologies

kryvasheyeu.pavel@gmail.comАпанасовичВ. В.ApanasovichV. V.

Апанасович Владимир Владимирович, доктор физикоматематических наук, профессор, первый проректор

Vladimir V. Apanasovich, Dr. Sci. (Phys.-Math.), Professor, First Vice-Rector

apanasovich@sbmt.byБелорусский государственный университетBelarusian State UniversityИнститут информационных технологий и бизнесадминистрированияInstitute of IT & Business Administration201906122018162717Copyright © Лисица Е.В., Яцков Н.Н., Скакун В.В., Кривошеев П.Д., Апанасович В.В., 20192019Лисица Е.В., Яцков Н.Н., Скакун В.В., Кривошеев П.Д., Апанасович В.В.Lisitsa Y.U., Yatskou M.M., Skakun V.V., Pavel D. Kryvasheyeu P.D., Apanasovich V.V.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://inf.grid.by/jour/article/view/474

Рассмотрены методы отбора информативных признаков для выделения геометрических признаков при описании ядер на люминесцентных изображениях раковых клеток. Выполнен обзор существующих геометрических признаков, который включает в себя как признаки формы, устойчивые к повороту и перемещению изображения, так и признаки расположения в пространстве. Для отбора наиболее информативных признаков использованы шесть методов: медианный, корреляционный с расчетом коэффициента корреляции по Пирсону, корреляционный с расчетом коэффициента корреляции по Спирмену, метод логистической регрессии, случайного леса с CART-деревьями и критерием Gini, случайного леса с CART-деревьями и критерием минимизации ошибки. В результате исследования из 59 признаков отобраны 11 наиболее информативных, выполнен анализ качества классификации с помощью метода случайного леса и рассчитаны временные затраты в зависимости от количества признаков для описания объектов. Для метода случайного леса использование 11 признаков является достаточным по точности классификации и позволяет снизить временные затраты в 2,3 раза.

The methods of geometric informative features selection of nuclei on fluorescent images of cancer cells are considered. During the survey, a review of existing geometric features was carried out, including both the signs of rotation resisted shape and displacement of the image, as well as signs of location in space. For the selection of characteristics, the methods were used: median, correlation with calculation of the Pearson correlation coefficient, correlation with calculation of the Spearman correlation coefficient, logistic regression model, random forest with CART trees and Gini criterion, random forest with CART trees and error minimization criterion. As a result of the investigation 11 characteristics were selected from 59 features, the quality of classification and time costs were calculated depending on the number of features for describing the objects. The use of 11 features is sufficient for the accuracy of classification as it allows to reduce time costs in 2,3 times.

корреляцияслучайный леслогическая регрессиямедианный методклассификацияcorrelationrandom forestlogistic regressionmedianclassificationReferencesStewart, B. World Cancer Report 2014 / B. Stewart, C. P. Wild. – Geneva : WHO Press, 2015. – 512 p.Stewart B., Wild C. P. World Cancer Report 2014. Geneva, WHO Press, 2015, 512 p.Программный пакет CellDataMiner для анализа люминесцентных изображений раковых клеток / Е. В. Лисица [и др.] // Информатика. – 2015. – № 4(48). – С. 73–84.Lisitsa Y. U., Yatskou M. М., Apanasovich V. V., Apanasovich T. V. Programmnyj paket CellDataMiner dlja analiza ljuminescentnyh izobrazhenij rakovyh kletok [The software package CellDataMiner for data mining of fluorescent images of cancer cells]. Informatics, 2015, no. 4(48), pp. 73–84 (in Russian).Spatial quantitative analysis of fluorescently labeled nuclear structures: problems, methods, pitfalls / O. Ronneberger [et al.] // Chromosome Research. – 2008. – No. 3. – P. 523–562.Ronneberger O., Baddeley D., Scheipl F., Verveer P. J., Burkhardt H., …, Joffe B. Spatial quantitative analysis of fluorescently labeled nuclear structures: problems, methods, pitfalls. Chromosome Research, 2008, no. 3, pp. 523–562.Supervised, unsupervised, and semi-supervised feature selection: a review on gene selection / J. C. Ang [et al.] // IEEE/ACM Transactions on Computational Biology and Bioinformatics. – 2016. – No. 5. – P. 971–989.Ang J. C., Mirzal A., Haron H., Hamed H. N. Supervised, unsupervised, and semi-supervised feature selection: a review on gene selection. IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2016, no. 5, pp. 971–989.Classifying ten types of major cancers based on reverse phase protein array profiles / P. W. Zhang [et al.] // PLoS One. – 2015. – No. 5. – P. 3–7.Zhang P. W., Chen L., Huang T., Zhang N., Kong X.Y., Cai Y. D. Classifying ten types of major cancers based on reverse phase protein array profiles. PLoS One, 2015, no. 5, pp. 3–7.Reverse phase protein array based tumor profiling identifies a biomarker signature for risk classification of hormone receptor-positive breast cancer / J. Sonntag [et al.] // Translational Proteomics. – 2014. – No. 2. – P. 52–59.Sonntag J., Bender C., Soons Z., Heyde S. von der, König R., …, Korf U. Reverse phase protein array based tumor profiling identifies a biomarker signature for risk classification of hormone receptor-positive breast cancer. Translational Proteomics, 2014, no. 2, pp. 52–59.Kaddi, C. Models for predicting stage in head and neck squamous cell carcinoma using proteomic and transcriptomic data / C. Kaddi, M. D.Wang // IEEE J. of Biomedical and Health Informatics. – 2017. – No. 1. – P. 246–253.Kaddi C., Wang M. D. Models for predicting stage in head and neck squamous cell carcinoma using proteomic and transcriptomic data. IEEE Journal of Biomedical and Health Informatics, 2017, no. 1, pp. 246–253.Immunosignature system for diagnosis of cancer / P. Stafford [et al.] // Proc. of the National Academy of Sciences of the United States of America. – 2014. – No. 30. – P. 3072–3080.Stafford P., Cichacz Z., Woodbury N. W., Johnston S. A. Immunosignature system for diagnosis of cancer. Proceedings of the National Academy of Sciences of the United States of America, 2014, no. 30, pp. 3072–3080.Nguyen, T. Modified AHP for gene selection and cancer classification using type-2 fuzzy logic / T. Nguyen, S. Nahavandi // IEEE Transactions on Fuzzy Systems. – 2016. – No. 2. – P. 273–287.Nguyen T., Nahavandi S. Modified AHP for gene selection and cancer classification using type-2 fuzzy logic. IEEE Transactions on Fuzzy Systems, 2016, no. 2, pp. 273–287.Hierarchical gene selection and genetic fuzzy system for cancer microarray data classification / T. Nguyen [et al.] // PloS One. – 2015. – No. 3.Nguyen T., Khosravi A., Creighton D., Nahavandi S. Hierarchical gene selection and genetic fuzzy system for cancer microarray data classification. PloS One, 2015, no. 3.3-phosphoadenosine 5-phosphosulfate (paps) synthases, naturally fragile enzymes specifically stabilized by nucleotide binding / J. Van den Boom [et al.] // J. of Biological Chemistry. – 2012. – No. 21. – P. 17645–17655.Boom J. Van den, Heider D., Martin S. R., Pastore A., Mueller J. W. 3-phosphoadenosine 5-phosphosulfate (paps) synthases, naturally fragile enzymes specifically stabilized by nucleotide binding. Journal of Biological Chemistry, 2012, no. 21, pp. 17645–17655.Insights into the classification of small GTPases / D. Heider [et al.] // Advances and Applications in Bioinformatics and Chemistry. – 2010. – No. 3. – P. 15–24.Heider D., Hauke S., Pyka M., Kessler D. Insights into the classification of small GTPases. Advances and Applications in Bioinformatics and Chemistry, 2010, no. 3, pp. 15–24.Data mining in the Life Sciences with Random Forest: a walk in the park or lost in the jungle? / W. G. Touw [et al.] // Briefings in Bioinformatics. – 2013. – No. 3. – P. 315–326.Touw W. G., Bayjanov J. R., Overmars L., Backus L., Boekhorst J., Wels M., Hijum van S. A. Data mining in the Life Sciences with Random Forest: a walk in the park or lost in the jungle? Briefings in Bioinformatics, 2013, no. 3, pp. 315–326.Improved Bevirimat resistance prediction by combination of structural and sequence-based classifiers / J. N. Dybowski [et al.] // BioData Mining. – 2011. – No. 4. – Р. 26–39.Dybowski J. N., Riemenschneider M., Hauke S., Pyka M., Verheyen J., Hoffmann D., Heider D. Improved Bevirimat resistance prediction by combination of structural and sequence-based classifiers. BioData Mining, 2011, no. 4, рр. 26–39.Exploiting HIV-1 protease and reverse transcriptase cross-resistance information for improved drug resistance prediction by means of multi-label classification / M. Riemenschneider [et al.] // BioData Mining. – 2016. – No. 9. – Р. 10–16.Riemenschneider M., Senge R., Neumann U., Hüllermeier E., Heider D. Exploiting HIV-1 protease and reverse transcriptase cross-resistance information for improved drug resistance prediction by means of multi-label classification. BioData Mining, 2016, no. 9, рр. 10–16.Expression profiling reveals novel pathways in the transformation of melanocytes to melanomas / H. Hoek [et al.] // Cancer Research. – 2004. – No. 15. – P. 5270–5282.Hoek H., Rimm D. L., Williams K. R., Zhao H., Ariyan S., …, Halaban R. Expression profiling reveals novel pathways in the transformation of melanocytes to melanomas. Cancer Research, 2004, no. 15, pp. 5270–5282.Quantitative analysis of estrogen receptor heterogeneity in breast cancer / G. G. Chung [et al.] // Laboratory Investigation. – 2007. – No. 7. – P. 662–669.Chung G. G., Zerkowski M. P., Ghosh S., Camp R. L., Rimm D. L. Quantitative analysis of estrogen receptor heterogeneity in breast cancer. Laboratory Investigation, 2007, no. 7, pp. 662–669.Camp, R. L. Automated subcellular localization and quantification of protein expression in tissue microarrays / R. L. Camp, G. G. Chung, D. L. Rimm // Nature Medicine. – 2002. – No. 11. – P. 1323–1327.Camp R. L., Chung G. G., Rimm D. L. Automated subcellular localization and quantification of protein expression in tissue microarrays. Nature Medicine, 2002, no. 11, pp. 1323–1327.Quantifying estrogen and progesterone receptor expression in breast cancer by digital imaging / M. K. Szesze [et al.] // J. of Histochemistry and Cytochemistry. – 2005. – No. 6. – P. 753–762.Szesze M. K., Crisman C. L., Crow L., McMullen S., Major J. M., …, Wasserman L. M. Quantifying estrogen and progesterone receptor expression in breast cancer by digital imaging. Journal of Histochemistry and Cytochemistry, 2005, no. 6, pp. 753–762.Имитационная модель трехканальных люминесцентных изображений популяций раковых клеток / Е. В. Лисица [и др.] // Журнал прикладной спектроскопии. – 2014. – № 6. – С. 907–913.Lisitsa Y. U., Yatskou M. М., Apanasovich V. V., Apanasovich T. V., Shitik M. M. Imitacionnaja model' trehkanal'nyh ljuminescentnyh izobrazhenij populjacij rakovyh kletok [Simulation model for three-channel luminescent images of cancer cell populations]. Zhurnal prikladnoj spektroskopii [Journal of Applied Spectroscopy], 2014, no. 6, pp. 907–913 (in Russian).Burger, W. Principles of Digital Image Processing: Core Algorithms / W. Burger, M. Burge. – London : Springer-Verlag, 2009. – 332 p.Burger W., Burge M. Principles of Digital Image Processing: Core Algorithms. London, SpringerVerlag, 2009, 332 p.Jähne, B. Digital Image Processing. Iss. 5 / B. Jähne. – Berlin, Heidelberg : Springer, 2002. – 585 p.Jähne B. Digital Image Processing. Iss. 5. Berlin, Heidelberg, Springer, 2002, 585 p.Reiss, Th. H. Recognizing Planar Objects using Invariant Image Features / Th. H. Reiss. – Berlin : Springer, 1993. – 186 p.Reiss Th. H. Recognizing Planar Objects using Invariant Image Features. Berlin, Springer, 1993, 186 p.Hu, M. K. Visual pattern recognition by moment invariants / M. K. Hu // IEEE Transactions on Information Theory. – 1962. – No. 8. – P. 179–187.Hu M. K. Visual pattern recognition by moment invariants. IEEE Transactions on Information Theory, 1962, no. 8, pp. 179–187.Neumann, U. EFS: an ensemble feature selection tool implemented as R-package and web-application / U. Neumann // BioData Mining. – 2017. – No. 10. – Р. 21–30.Neumann U. EFS: an ensemble feature selection tool implemented as R-package and web-application. BioData Mining, 2017, no. 10, рр. 21–30.Bauer, D. F. Constructing confidence sets using rank statistics / D. F. Bauer // J. of the American Statistical Association. – 1972. – No. 67. – P. 687–690.Bauer D. F. Constructing confidence sets using rank statistics. Journal of the American Statistical Association, 1972, no. 67, pp. 687–690.Yu, L. Efficient feature selection via analysis of relevance and redundancy / L. Yu // J. of Machine Learning Research. – 2004. – No. 5. – P. 1205–1224.Yu L. Efficient feature selection via analysis of relevance and redundancy. Journal of Machine Learning Research, 2004, no. 5, pp. 1205–1224.Suzuki, N. Developing landscape habitat models for rare amphibians with small geographic ranges: a case study of Siskiyou Mountains salamanders in the western USA / N. Suzuki, D. H. Olson, E. C. Reilly // J. of Machine Learning Research. – 2008. – No. 17. – P. 2197–2218.Suzuki N., Olson D. H., Reilly E. C. Developing landscape habitat models for rare amphibians with small geographic ranges: a case study of Siskiyou Mountains salamanders in the western USA. Journal of Machine Learning Research, 2008, no. 17, pp. 2197–2218.Novel methods improve prediction of species distributions from occurrence data / J. Elith [et al.] // J. of Space and Time in Ecology. – 2006. – No. 29. – P. 129–151.Elith J., Graham C. H., Anderson R. P., Dudík M., Ferrier S., …, Zimmermann N. E. Novel methods improve prediction of species distributions from occurrence data. Journal of Space and Time in Ecology, 2006, no. 29, pp. 129–151.Yu, L. Efficient feature selection via analysis of relevance and redundancy / L. Yu, H. Liu // J. of Machine Learning Research. – 2004. – No. 5. – P. 1205–1224.Yu L., Liu H. Efficient feature selection via analysis of relevance and redundancy. Journal of Machine Learning Research, 2004, no. 5, pp. 1205–1224.Compensation of feature selection biases accompanied with improved predictive performance for binary classification by using a novel ensemble feature selection approach / U. Neumann [et al.] // BioData Mining. – 2016. – No. 9. – P. 36–50.Neumann U., Riemenschneider M., Sowa J.-P., Baars T., Kälsch J., Canbay A., Heider D. Compensation of feature selection biases accompanied with improved predictive performance for binary classification by using a novel ensemble feature selection approach. BioData Mining, 2016, no. 9, pp. 36–50.Breiman, L. Random forests / L. Breiman // Machine Learning. – 2001. – No. 5. – P. 5–32.Breiman L. Random forests. Machine Learning, 2001, no. 5, pp. 5–32. 33. Liu J., Lin Y., Lin M., Wu S., Zhang J. Feature selection based on quality of information. Neurocomputing, 2017, no. 225, pp. 11–22.Feature selection based on quality of information / J. Liu [et al.] // Neurocomputing. – 2017. – No. 225. – P. 11–22.Montañés E., Fernández J., Díaz I., Combarro E. F., Ranilla J. Measures of rule quality for feature selection in text categorization. Advances in Intelligent Data Analysis V, 2003, no. 225, pp. 589–598.Measures of rule quality for feature selection in text categorization / E. Montañés [et al.] // Advances in Intelligent Data Analysis V. – 2003. – No. 225. – P. 589–598.Pedregosa F., Varoquaux G., Gramfort A., Michel V., Thirion B., …, Duchesnay É. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 2011, no. 12, pp. 2825–2830.Scikit-learn: Machine learning in Python / F. Pedregosa [et al.] // J. of Machine Learning Research. – 2011. – No. 12. – P. 2825–2830.Scikit-learn: Machine learning in Python / F. Pedregosa [et al.] // J. of Machine Learning Research. – 2011. – No. 12. – P. 2825–2830.

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