V.V. Khryashchev1

1 P.G. Demidov Yaroslavl State University (Yaroslavl, Russia)

Algorithms and methods for analyzing video images in medicine are used in such a promising field of diagnostics as colonoscopic examinations. Such an analysis makes it possible to detect and classify dangerous pathologies, including oncological ones, at an early stage. One of the markers of such pathologies are colon polyps – abnormal tissue growths protruding above the mucous membrane. They are often harbingers of fairly common colorectal cancer. To investigate neural network algorithms for segmentation of intestinal polyps images based on deep machine learning methods. For training and testing of algorithms, both the open international database of polyp images Kvasir-SEG and the original database of images OnkoYar-SEG, collected in the endoscopic department of the Yaroslavl Regional Clinical Oncology Hospital, were used. Studies show that the SSformer neural network algorithm with a value of 0.96 according to the Dice metric shows the best result among the studied algorithms. Its advantage over the classical neural network algorithm U-Net is 15%, which is a significant improvement in the accuracy of polyp segmentation. Additional testing using the OnkoYar-SEG database confirmed the observed advantage of the algorithm based on the SSformer neural network architecture. The proposed algorithm can be used as a basis for a neural network system for analy­zing polyps during colonoscopic examinations.

Khryashchev V.V. Image segmentation of colorectal polyps during colonoscopy using neural networks. Biomedicine Radioengi­neering. 2023. V. 26. № 4. P. 66–72. DOI: https://doi.org/10.18127/j15604136-202302-07 (In Russian)

1. Palevskaya S.A., Korotkevich A.G. Endoskopiya zheludochno-kishechnogo trakta. M.: GEOTAR-Media. 2020.
2. Kashin S.V., Nekhajkova N.V., Zav'yalov D.V., Vidyaeva N.S., Belova A.N. Skrining kolorektal'nogo raka: obshchaya situaciya v mire i rekomendovannye standarty kachestva kolonoskopii. Dokazatel'naya gastroenterologiya. 2017. T. 6. № 4. S. 32–52.
3. Nikonov E.L., Aksenov V.A., Kashin S.V., Nekhajkova N.V. Mezhdunarodnyj opyt skrininga kolorektal'nogo raka. Dokazatel'naya gastroenterologiya. 2017. T. 6. № 3. S. 30–35.
4. Lebedev A.A., Hryashchev V.V., Srednyakova A.S., Kazina E.M. Razrabotka algoritma detektirovaniya polipov na endoskopicheskih izobrazheniyah c ispol'zovaniem svertochnyh nejronnyh setej. Cifrovaya obrabotka signalov. 2021. № 2. S. 55–60.
5. Lebedev A.A. Issledovanie nejrosetevyh algoritmov obnaruzheniya ob"ektov na videoizobrazheniyah v medicinskih sistemah prikladnogo televideniya: diss. kand. tekhn. nauk. YAroslavl'. 2022.
6. Karpov O.E., Hramov A.E. Informacionnye tekhnologii, vychislitel'nye sistemy i iskusstvennyj intellekt v medicine. M.: DPK Press. 2022.
7. Rangayyan R.M. Biomedical Image Analysis. CRC Press. Boca Raton. 2005. 1306 p.
8. Stepanova O.A. Algoritmy raspoznavaniya gistologicheskih i endoskopicheskih izobrazhenij dlya sistem prikladnogo televideniya v medicine: diss. … kand. tekhn. nauk. YAroslavl'. 2019.
9. Masci J., Meier U., Cires D., Schmidhuber J. Stacked convolutional auto-encoders for hierarchical feature extraction. In International Conference on Artificial Neural Networks. 2011. Springer. R. 52–59.
10. Long J., Shelhamer E., Darrell T. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition. 2015. R. 3431–3440.
11. Badrinarayanan V., Kendall A., Cipolla R. Segnet: A deep convolutional encoder-decoder architecture for image segmentation. IEEE transactions on pattern analysis and machine intelligence. 2017. 39 (12). R. 2481–2495.
12. Ronneberger O., Fischer P., Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation. Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. Lecture Notes in Computer Science. Springer. 2015. vol. 9351. R. 234–241.
13. Zhou Z., Siddiquee M., Tajbakhsh N., Liang J. UNet++: A Nested U-Net Architecture for Medical Image Segmentation. arXiv:1807.10165v1 [cs.CV] 18 Jul 2018.
14. Jha D., Smedsrud P., Riegler M., Johansen D., Lange T. et. al. ResUNet++: An Advanced Architecture for Medical Image Segmenta­tion. in 21st IEEE International Symposium on Multimedia. 2019.
15. Azad R., Heidari M., Shariatnia M., Aghdam E., Karimijafarbigloo S. et. al. TransDeepLab: Convolution-Free Transformer-based DeepLab v3+ for Medical Image Segmentation. 2022. arXiv:2208.00713v1 [eess.IV] 1 Aug 2022.
16. Shi W., Xu J., Gao P. SSformer: A Lightweight Transformer for Semantic Segmentation.  arXiv:2208.02034v1 [cs.CV] 3 Aug 2022.
17. Jha D., Smedsrud P., Riegler M., Halvorsen P., Lange T. et. al. Kvasir-seg: A Segmented Polyp Dataset. Proc. of International Con-ference on Multimedia Modeling. 2020. R. 451–462.
18. Jha D., Ali S., Tomar N.K., Johansen H., Johansen D. et. al. Real-Time Polyp Detection, Localization and Segmentation in Colonos-copy Using Deep Learning. Computer Vision and Pattern Recognition. 2021. R. 40496–40510.
19. Image Polygonal Annotation with Python – LabelMe (https://github.com/wkentaro/labelme).
20. Carass A., Roy S., Gherman A., Reinhold J., Jesson A. et. al. Evaluating White Matter Lesion Segmentations with Refined Sørensen-Dice Analysis. Sci Rep. 2020 May 19;10(1):8242. doi: 10.1038/s41598-020-64803