sapiСистемный анализ и прикладная информатика«System analysis and applied information science»2309-49232414-0481Belarusian National Technical Universitysapi-106Research ArticleОбработка информации и принятие решенийData processing and decision–makingМЕТОДИКА ПОДСЧЕТА ЧИСЛА ЯДЕР КЛЕТОК НА МЕДИЦИНСКИХ ГИСТОЛОГИЧЕСКИХ ИЗОБРАЖЕНИЯХAN APPROACH TO CELL NUCLEI COUNTING IN HISTOLOGICAL IMAGE ANALYSISЛукашевичМ. М.LukashevichM. M.

Лукашевич Марина Михайловна, кандидат технических наук, доцент кафедры ЭВМ

valerys@newman.bas-net.byСтаровойтовВ. В.StarovoitovV. V.

Старовойтов Валерий Васильевич, доктор технических наук, профессор, главный научный сотрудник

valerys@newman.bas-net.byBelarusian State University of Informatics and RadioelectronicsBelarusUnited Institute of Informatics Problems of the NAS of BelarusBelarus201626072016023742Copyright © Лукашевич М.М., Старовойтов В.В., 20162016Лукашевич М.М., Старовойтов В.В.Lukashevich M.M., Starovoitov 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://sapi.bntu.by/jour/article/view/106

В статье исследуется методика автоматического подсчета числа ядер клеток на гистологических изображениях. Эта операция широко применяется при диагностике различных заболеваний и морфологическом анализе клеток. В связи с этим, процедура автоматического подсчета числа ядер клеток является ключевым этапом в системах микроскопического анализа медицинских изображений гистологических препаратов. Основной целью работы была разработка эффективной схемы автоматического подсчета ядер клеток на основе современных методов обработки изображений: направленной фильтрации, адаптивной бинаризации изображений и математической морфологии. В отличие от известных исследований, представленный подход не предусматривает сегментацию ядер клеток на изображении, а лишь предполагает их обнаружение и подсчет их количества. Это позволяет избежать сложных алгоритмических вычислений и обеспечивает хорошую точность подсчета ядер клеток.

В работе описан ряд экспериментов, выполненных для оценки эффективности предложенной методики с использованием доступной в интернете тестовый базы медицинских гистологических изображений. Определены критичные параметры алгоритмов, настраиваемые на каждом этапе анализа изображений. Для каждого параметра определен интервал тестируемых значений, а затем реализована процедура выбора не только оптимальных значений каждого параметра, но их из взаимная комбинация, на основе общепринятых количественных оценок точности (Precision) и полноты (Recall). Полученные результаты сравнивались с последними достижениями в данной области и показали приемлемый уровень точности предложенной методики. Прототип программного обеспечения, разработанного в рамках проведенного исследования, можно рассматривать как автоматический инструмент для анализа ядер клеток. Разработанный подход может быть адаптирован к различным задачам анализа ядер клеток различных органов.

In the paper a method of automatical counting the number of cell nuclei in histological images is studied. This operation is commonly used in the diagnostics of various diseases and morphological analysis of cells. In this connection, the procedure of automatical count the number of cell nuclei is a key step in the systems of medical imaging microscopic analysis of histological preparations. The main aim of our work was to develop an efficient scheme of automatic counting cell nuclei based on advanced image processing methods: directional filtering, adaptive image binarization and mathematical morphology. Unlike prior research, the presented approach does not provide segmentation of cell nuclei in the image, but only requires to detect them and count their number. This avoids complex algorithmic calculations and provides good accuracy of counting cell nuclei.

The paper describes a series of experiments conducted to assess the effectiveness of the proposed method using the available online database of medical test histological images. Critical parameters defined algorithms, configurable at each stage of image analysis. For each parameter we have defined value ranges, and then realized a selection of optimal values for every parameter and a mutual combination of them. It is based on generally accepted quantitative measures of precision and recall. The results were compared with the state-of-art investigations in this field and demonstrated an acceptable level of accuracy of the proposed method. The software prototype developed during the study can be regarded as an automatic tool for analysis of cell nuclei. The presented approach can be adapted to various problems of analysis of cell nuclei of various organs.

изображения гистологических препаратовобработка и анализ цифровых изображенийбинаризацияРабота частично выполнена в рамках гранта фонда фундаментальных исследований № Ф15ЛИТ-031.ReferencesMethods For Nuclei Detection, Segmentation, and Classification in Digital Histopathology: A Review – Current Status and Future Potential / H. Irshad [et. al] // IEEE Reviews in Biomedical Engineering, 2014. − Vol. 7. − P. 97−114.Methods For Nuclei Detection, Segmentation, and Classification in Digital Histopathology: A Review – Current Status and Future Potential / H. Irshad [et. al] // IEEE Reviews in Biomedical Engineering, 2014. − Vol. 7. − P. 97−114.Recent Advances in Morphological Cell Image Analysis / S. Chen [et. al] // Computational and Mathematical Methods in Medicine, 2012. – 10 p.Recent Advances in Morphological Cell Image Analysis / S. Chen [et. al] // Computational and Mathematical Methods in Medicine, 2012. – 10 p.Jung, C. Impact of the Accuracy of Automatic Segmentation of Cell Nuclei Clusters on Classification of Cell Nuclei Clusters on Classification of Thyroid Follicular Lesions / C. Jung, C. Kim // Cytometry. Part A, 2014. – P. 709–719.Jung, C. Impact of the Accuracy of Automatic Segmentation of Cell Nuclei Clusters on Classification of Cell Nuclei Clusters on Classification of Thyroid Follicular Lesions / C. Jung, C. Kim // Cytometry. Part A, 2014. – P. 709–719.Multi-resolution Approach for Combining Visual Information using Nuclei Segmentation and Classification in Histopathological Images / H. Saharma [et. al] // Proceedings of the 10th International Conference on Computer Vision, Theory and Applications, 2015. – P. 37−46.Multi-resolution Approach for Combining Visual Information using Nuclei Segmentation and Classification in Histopathological Images / H. Saharma [et. al] // Proceedings of the 10th International Conference on Computer Vision, Theory and Applications, 2015. – P. 37−46.Alilou, M. Segmentation of cell nuclei in heterogeneous microscopy images: A resalable templates approach / M. Alilou, V. Kovalev, V. Taimouri // Computerized Medical Imaging and Graphics, 2013. – Vol. 37. – P. 488−499.Alilou, M. Segmentation of cell nuclei in heterogeneous microscopy images: A resalable templates approach / M. Alilou, V. Kovalev, V. Taimouri // Computerized Medical Imaging and Graphics, 2013. – Vol. 37. – P. 488−499.Kowal, M. Nuclei Segmentation for Computer-Aded Diagmosis of Breast Cancer / M. Kowal, P. Filipczuk // Int. Journal Appl. Math. Comput. Science, 2014. – Vol. 24. – No. 1. – P. 19−31.Kowal, M. Nuclei Segmentation for Computer-Aded Diagmosis of Breast Cancer / M. Kowal, P. Filipczuk // Int. Journal Appl. Math. Comput. Science, 2014. – Vol. 24. – No. 1. – P. 19−31.Detection and segmentation of cell nuclei in virtual microscopy images: a minimums-model approach / S. Wienert [et al] // National Scientific Reports, 2012. – 2:503.Detection and segmentation of cell nuclei in virtual microscopy images: a minimums-model approach / S. Wienert [et al] // National Scientific Reports, 2012. – 2:503.White Blood Cell Segmentation by Color-Space-Based K-Means Clustering / C. Zang [et. al] // Sensors, 2014. Vol. – 14. – P. 16128–16147.White Blood Cell Segmentation by Color-Space-Based K-Means Clustering / C. Zang [et. al] // Sensors, 2014. Vol. – 14. – P. 16128–16147.Region-based progressive localization of cell nuclei in microscopic images with data adaptive modeling / Y. Song [et. al] // BMC Bioinformatics. – 2013. Vol. 14. № 1. – P. 173–180.Region-based progressive localization of cell nuclei in microscopic images with data adaptive modeling / Y. Song [et. al] // BMC Bioinformatics. – 2013. Vol. 14. № 1. – P. 173–180.Coelho, L. P. Nuclear segmentation in microscope cell images: a hand-segmented dataset and comparison of algorithms / L. P. Coelho, A. Shariff, R. F., Murphy // Proc. of the IEEE International Symposium Biomedical Imaging, 2009. – P. 518−521.Coelho, L. P. Nuclear segmentation in microscope cell images: a hand-segmented dataset and comparison of algorithms / L. P. Coelho, A. Shariff, R. F., Murphy // Proc. of the IEEE International Symposium Biomedical Imaging, 2009. – P. 518−521.Wavelet-Based Multiscale Texture Segmentation: Application to Stromal Compartment Characterization on Virtual Slides / N. Signolle [et. al] // Signal Processing, 2010. – Vol. 90. – № 8. – P. 2412–2422.Wavelet-Based Multiscale Texture Segmentation: Application to Stromal Compartment Characterization on Virtual Slides / N. Signolle [et. al] // Signal Processing, 2010. – Vol. 90. – № 8. – P. 2412–2422.Segmentation of cytological image using color and mathematical morphology / O. Lezoray [et. al] // Acta Stereologica, 1999. – 18. – P. 1−14.Segmentation of cytological image using color and mathematical morphology / O. Lezoray [et. al] // Acta Stereologica, 1999. – 18. – P. 1−14.An Image Analysis-Based Approach for Automated Counting of Cancer Cell Nuclei / C. G. Loukas [et. al] // Cytomettry. Part A, 2003. – P. 30−42.An Image Analysis-Based Approach for Automated Counting of Cancer Cell Nuclei / C. G. Loukas [et. al] // Cytomettry. Part A, 2003. – P. 30−42.Improved automatic detection and segmentation of cell nuclei in histopathology images / Y. Al-Kofahi [et. al] // IEEE Trans. on Biomedical Engineering, 2010. – Vol. 57. – № 4. – P. 841−852.Improved automatic detection and segmentation of cell nuclei in histopathology images / Y. Al-Kofahi [et. al] // IEEE Trans. on Biomedical Engineering, 2010. – Vol. 57. – № 4. – P. 841−852.Cell-based quantification of molecular biomarkers in histopathology specimens / Y. Al-Kofahi [et. al] // Histopathology, 2011. – 59(1) – P. 40−54.Cell-based quantification of molecular biomarkers in histopathology specimens / Y. Al-Kofahi [et. al] // Histopathology, 2011. – 59(1) – P. 40−54.Image Smoothing via L0 Gradient Minimization / L. Xu [et. al] // ACM Transactions on Graphics. – December 2011. – Vol. 30. – No. 6. – article 174.Image Smoothing via L0 Gradient Minimization / L. Xu [et. al] // ACM Transactions on Graphics. – December 2011. – Vol. 30. – No. 6. – article 174.He, K. Guided image filtering / K. He, J. Sun, X. Tang // Pattern Analysis and Machine Intelligence, IEEE Transactions on. – 2013. – Т. 35. – №. 6. – С. 1397–1409.He, K. Guided image filtering / K. He, J. Sun, X. Tang // Pattern Analysis and Machine Intelligence, IEEE Transactions on. – 2013. – Т. 35. – №. 6. – С. 1397–1409.Wan, S. J. Variance‐based color image quantization for frame buffer display / S. J. Wan, P. Prusinkiewicz, S. K. M. Wong // Color Research & Application. – 1990. – Т. 15. – №. 1. – С. 52–58.Wan, S. J. Variance‐based color image quantization for frame buffer display / S. J. Wan, P. Prusinkiewicz, S. K. M. Wong // Color Research & Application. – 1990. – Т. 15. – №. 1. – С. 52–58.Sauvola, J. Adaptive document image binarization / J. Sauvola, M. Pietikainen // Pattern Recognition, 2000. – Vol. 33. – P. 225−236.Sauvola, J. Adaptive document image binarization / J. Sauvola, M. Pietikainen // Pattern Recognition, 2000. – Vol. 33. – P. 225−236.Shafait, F. Efficient implementation of local adaptive thresholding techniques using integral images / F. Shafait, D. Keysers, T. M. Breuel // Document Recognition and Retrieval XV. – 2008.Shafait, F. Efficient implementation of local adaptive thresholding techniques using integral images / F. Shafait, D. Keysers, T. M. Breuel // Document Recognition and Retrieval XV. – 2008.Stathis, P. An Evaluation Technique for Binarization Algorithms / P. Stathis, E. Kavallieratou, N. Papamarkos // Journal of Universal Computer Science, 2008. – Vol. 14. – No. 18. – P. 3011−3030.Stathis, P. An Evaluation Technique for Binarization Algorithms / P. Stathis, E. Kavallieratou, N. Papamarkos // Journal of Universal Computer Science, 2008. – Vol. 14. – No. 18. – P. 3011−3030.

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