Cervical cell extraction network based on optimized yolo

Nengkai Wu; Dongyao Jia; Chuanwang Zhang; Ziqi Li; Nengkai Wu; Dongyao Jia; Chuanwang Zhang; Ziqi Li

doi:10.3934/mbe.2023111

Mathematical Biosciences and Engineering

2023, Volume 20, Issue 2: 2364-2381. doi: 10.3934/mbe.2023111

Previous Article Next Article

Research article

Cervical cell extraction network based on optimized yolo

Beijing Jiaotong University, School of Electronics and Information Engineering, No. 3 Shangyuancun Haidian District, Beijing, China, 100044

Received: 26 September 2022 Revised: 31 October 2022 Accepted: 09 November 2022 Published: 18 November 2022

Early screening for cervical cancer is a common form of cancer prevention. In the microscopic images of cervical cells, the number of abnormal cells is small, and some abnormal cells are heavily stacked. How to solve the segmentation of highly overlapping cells and realize the identification of single cells from overlapping cells is still a heavy task. Therefore, this paper proposes an object detection algorithm of Cell_yolo to effectively and accurately segment overlapping cells. Cell_yolo adopts a simplified network structure and improves the maximum pooling operation, so that the information of the image is preserved to the greatest extent during the model pooling process. Aiming at the characteristics of many overlapping cells in cervical cell images, a non-maximum suppression method of center distance is proposed to prevent the overlapping cell detection frame from being deleted by mistake. At the same time, the loss function is improved and the focus loss function is added to alleviate the imbalance of positive and negative samples in the training process. Experiments are conducted on a private dataset (BJTUCELL). Experiments have verified that the Cell_yolo model has the advantages of low computational complexity and high detection accuracy, and it is superior to common network models such as YOLOv4 and Faster_RCNN.
- cervical cells,
- yolo,
- overlapping cell,
- deep learning,
- object detection
Citation: Nengkai Wu, Dongyao Jia, Chuanwang Zhang, Ziqi Li. Cervical cell extraction network based on optimized yolo[J]. Mathematical Biosciences and Engineering, 2023, 20(2): 2364-2381. doi: 10.3934/mbe.2023111

Related Papers:

Abstract

Early screening for cervical cancer is a common form of cancer prevention. In the microscopic images of cervical cells, the number of abnormal cells is small, and some abnormal cells are heavily stacked. How to solve the segmentation of highly overlapping cells and realize the identification of single cells from overlapping cells is still a heavy task. Therefore, this paper proposes an object detection algorithm of Cell_yolo to effectively and accurately segment overlapping cells. Cell_yolo adopts a simplified network structure and improves the maximum pooling operation, so that the information of the image is preserved to the greatest extent during the model pooling process. Aiming at the characteristics of many overlapping cells in cervical cell images, a non-maximum suppression method of center distance is proposed to prevent the overlapping cell detection frame from being deleted by mistake. At the same time, the loss function is improved and the focus loss function is added to alleviate the imbalance of positive and negative samples in the training process. Experiments are conducted on a private dataset (BJTUCELL). Experiments have verified that the Cell_yolo model has the advantages of low computational complexity and high detection accuracy, and it is superior to common network models such as YOLOv4 and Faster_RCNN.

References

[1]	R. Elakkiya, K. S. S. Teja, L. J. Deborah, C. Bisogni, C. Medaglia, Imaging based cervical cancer diagnostics using small object detection-generative adversarial networks, Mult. Tools Appl., 81 (2022), 191–207. https://doi.org/10.1007/s11042-021-10627-3 doi: 10.1007/s11042-021-10627-3
[2]	J. C. Davies-Oliveira, M. A. Smith, S. Grover, K. Canfell, E. J. Crosbie, Eliminating cervical cancer: progress and challenges for high-income countries, Clin. Oncol., 33 (2021), 550–559. https://doi.org/10.1016/j.clon.2021.06.013 doi: 10.1016/j.clon.2021.06.013
[3]	M. E. Plissiti, C. Nikou, A Review of Automated Techniques for Cervical Cell Image Analysis and Classification, Springer Netherlands, 4 (2013), 1–18. https://doi.org/10.1007/978-94-007-4270-3_1 doi: 10.1007/978-94-007-4270-3_1
[4]	M. E. Plissiti, C. Nikou, Overlapping Cell Nuclei Segmentation Using a Spatially Adaptive Active Physical Model, IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society, 21 (2012), 4568–580. https://doi.org/10.1109/TIP.2012.2206041 doi: 10.1109/TIP.2012.2206041
[5]	N. M. Harandi, S. Sadri, N. A. Moghaddam, R. Amirfattahi, An Automated Method for Segmentation of Epithelial Cervical Cells in Images of ThinPrep, J. Med. Syst., 34 (2010), 1043–1058. https://doi.org/10.1007/s10916-009-9323-4 doi: 10.1007/s10916-009-9323-4
[6]	A. Genctav, S. Aksoy, S. Onder, Unsupervised segmentation and classification of cervical cell images, Pattern Recogn., 45 (2012), 4151–4168. https://doi.org/10.1016/j.patcog.2012.05.006 doi: 10.1016/j.patcog.2012.05.006
[7]	A. Kale, S. Aksoy, Segmentation of cervical cell images, 2010 20th International Conference on Pattern Recognition, IEEE, (2010), 2399–2402. https://doi.org/10.1109/ICPR.2010.587 doi: 10.1109/ICPR.2010.587
[8]	T. Chankong, N. Theera-Umpon, S. Auephanwiriyakul, Automatic cervical cell segmentation and classification in Pap smears, Computer Meth. Progr. Biomed., 113 (2014), 539–556. https://doi.org/10.1016/j.cmpb.2013.12.012 doi: 10.1016/j.cmpb.2013.12.012
[9]	H. Lee, J. Kim, Segmentation of overlapping cervical cells in microscopic images with superpixel partitioning and cell-wise contour Refinement, 29th IEEE Conference on Computer Vision and Pattern Recognition, (2016), 1367–1373. https://doi.org/10.1109/CVPRW.2016.172 doi: 10.1109/CVPRW.2016.172
[10]	B. Dong, L. Jia, Y. Wang, J. Li, G. Yang, An improved watershed algorithm based on k-medoids in cervical cancer images, 2019 IEEE International Conferences on Ubiquitous Computing & Communications (IUCC) and Data Science and Computational Intelligence (DSCI) and Smart Computing, Networking and Services (SmartCNS), IEEE, (2019), 190–195. https://doi.org/10.1109/IUCC/DSCI/SmartCNS.2019.00060 doi: 10.1109/IUCC/DSCI/SmartCNS.2019.00060
[11]	C. Jung, C. Kim, S. W. Chae, S. Oh, Unsupervised Segmentation of Overlapped Nuclei Using Bayesian Classification, IEEE Transact. Biomed. Eng., 57 (2010), 2825–2832. https://doi.org/10.1109/tbme.2010.2060486 doi: 10.1109/tbme.2010.2060486
[12]	D. N. Diniz, M. T. Rezende, A. G. C. Bianchi, C. M. Carneiro, D. M. Ushizima, F. N. S. de Medeiros, M. J. F. Souza, A hierarchical feature-based methodology to perform cervical cancer classification, Appl. Sciences-Basel, 11 (2021), 4091. https://doi.org/10.3390/app11094091 doi: 10.3390/app11094091
[13]	J. Long, E. Shelhamer, T. Darrell, Fully convolutional networks for semantic segmentation, The IEEE conference on computer vision and pattern recognition, 39 (2017), 640–651. https://doi.org/10.1109/TPAMI.2016.2572683 doi: 10.1109/TPAMI.2016.2572683
[14]	V. Badrinarayanan, A. Kendall, R. Cipolla, SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation, IEEE Transact. Pattern Anal. Mach. Intell., 39 (2017), 2481–2495. https://doi.org/10.1109/TPAMI.2016.2644615 doi: 10.1109/TPAMI.2016.2644615
[15]	E. Giacomello, D. Loiacono, L. Mainardi, Brain MRI Tumor Segmentation with Adversarial Networks, 2020 International Joint Conference on Neural Networks (IJCNN), IEEE, (2020), 1–8. https://doi.org/10.1109/IJCNN48605.2020.9207220 doi: 10.1109/IJCNN48605.2020.9207220
[16]	D. Yang, D. Xu, S. K. Zhou, B. Georgescu, M. Q. Chen, D. Comaniciu, Automatic liver segmentation using an adversarial image-to-image network, International Conference on Medical Image Computing and Computer-Assisted Intervention, (2017), 507–515. https://doi.org/10.1007/978-3-319-66179-7_58 doi: 10.1007/978-3-319-66179-7_58
[17]	R. Krithiga, P. Geetha, Breast cancer detection, segmentation and classification on histopathology images analysis: A systematic review, Arch. Comput. Methods Eng., 28 (2021), 2607–2619. https://doi.org/10.1007/s11831-020-09470-w doi: 10.1007/s11831-020-09470-w
[18]	O. Ronneberger, P. Fischer, T. Brox, U-net: Convolutional networks for biomedical image segmentation, International Conference on Medical image computing and computer-assisted intervention. Springer. Cham., (2015), 234–241. https://doi.org/10.48550/arXiv.1505.04597 doi: 10.48550/arXiv.1505.04597
[19]	X. Y. Li, L. L. Shen, cC-GAN: A robust transfer-learning framework for HEp-2 specimen image segmentation, IEEE Access, (2018), 14048–14058. https://doi.org/10.1109/access.2018.2808938 doi: 10.1109/access.2018.2808938
[20]	Y. Nambu, T. Mariya, S. Shinkai, M. Umemoto, H. Asanuma, I. Sato, et al., A screening assistance system for cervical cytology of squamous cell atypia based on a two-step combined CNN algorithm with label smoothing, Cancer Med., 11 (2022), 520–529. https://doi.org/10.1002/cam4.4460 doi: 10.1002/cam4.4460
[21]	Z. Q. Xing, X. Chen, F. Q. Pang, DD-YOLO: An object detection method combining knowledge distillation and Differentiable Architecture Search, IET Computer Vision, 16 (2022), 418–430. https://doi.org/10.1049/cvi2.12097 doi: 10.1049/cvi2.12097
[22]	L. C. Jiao, F. Zhang, F. Liu, S. Y. Yang, L. L. Li, Z. X. Feng, et al., A survey of deep learning-based object detection, IEEE Access, 7 (2019), 12883–128868. https://doi.org/10.1109/ACCESS.2019.2939201 doi: 10.1109/ACCESS.2019.2939201
[23]	A. Bochkovskiy, C. Y. Wang, H. Liao, YOLOv4: Optimal speed and accuracy of object detection, Computer Sci., (2020). https://doi.org/10.48550/arXiv.2004.10934 doi: 10.48550/arXiv.2004.10934
[24]	S. Liu, L. Qi, H. F. Qin, J. P. Shi, J. Y. Jia, Path aggregation network for instance segmentation, 31st IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), (2018), 8759–8768. https://doi.org/10.1109/CVPR.2018.00913 doi: 10.1109/CVPR.2018.00913

Reader Comments

Your name:*

Email:*
© 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)