COMPUTER VISION MODEL COMPARISON
https://doi.org/10.55452/1998-6688-2021-18-3-75-82
Abstract
The use of machine learning in the medical field is one of the most difficult and thoroughly unsolved problems. Currently, there are many different algorithms for solving problems in the field of diagnostics and segmentation of biomedical images. Researchers are often faced with the challenge of choosing the best method to apply towards their data. We conducted the empirical research and compared 5 algorithms that able to detect anomalies in the medical images: RCNN, Fast-RCNN, Faster-RCNN, Mask R CNN, U-Net, and Residual Neural Network. The advantages of automatic processing of the medical images are apparent: doctors get a convenient software tool that allows them to diagnose the disease faster and reduce possible errors. The task is to study and then select algorithms for further testing on the actual data. The selection and study of algorithms were based on articles describing the architecture and application of computer vision algorithms.
About the Authors
D. Nam
Kazakh-British Technical University
Kazakhstan
Kazakhstan
Nam Diana
050000, Almaty
T. Savina
Kazakh-British Technical University
Kazakhstan
Kazakhstan
Savina Tamara
050000, Almaty
References
1. R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierar-chies for accurate object detection and semantic segmentation,”Proceed-ings of the IEEE Computer Society Conference on Computer Vision andPattern Recognition, pp. 580–587, 2014.
2. R. Girshick, “Fast R-CNN,”Proceedings of the IEEE International Conference on Computer Vision, vol. 2015 Inter, pp. 1440–1448, 2015
3. S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-timeobject detection with region proposal networks,” 2016.
4. K. He, G. Gkioxari, P. Doll ár, and R. Girshick, “Mask R-CNN,”IEEETransactions on Pattern Analysis and Machine Intelligence, vol. 42, no. 2,pp. 386–397, 2020.
5. J. W. Johnson, “Adapting Mask-RCNN for automatic nucleus segmentation,” arXiv, pp. 1–7, 2018.
6. Image set BBBC038v1, available from the Broad Bioimage Benchmark Collection [Caicedo et al., Nature Methods, 2019]
7. S. Liu, L. Qi, H. Qin, J. Shi, and J. Jia, “Path aggregation network forinstance segmentation,” 2018
8. E. Kopelowitz and G. Engelhard, “Lung nodules detection and segmen-tation using 3d mask-rcnn,” 2019
9. C. Zhao, J. Han, Y. Jia, and F. Gou, “Lung nodule detection via3D U-net and contextual convolutional neural network,”Proceedings -2018 International Conference on Networking and Network Applications,NaNA 2018, pp. 356–361, 2019. A. Pak, A. Ziyaden, K. Tukeshev, A. Jaxylykova, and D. Abdullina,“Comparative analysis of deep learning methods of detection of diabeticretinopathy,”Cogent Engineering, vol. 7, no. 1, p. 1805144, 2020.
10. O. Ronneberger, P. Fischer, T. Brox, “U-net: Convolutional networks for biomedical image segmentation”, vol. 9351, pp.234-241, 2015.
11. M. Zhou, M. Rahman, N. Siddiquee, Tajbakhsh, and J. Liang, Unet++, “A nested u-net architecture for medical image segmentation”, Springer International Publishing, 2018, vol. 11045 LNCS. [Online].
12. K.He, X.Zhang, S. Ren, “Deep residual learning for image recognition”, Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2016-decem, p.770-778, 2016.
13. Romera-Paredes and P. H. S. Torr, “Recurrent instance segmentation,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9910 LNCS, pp. 312–329, 2016.
14. O. Vinyals, Q.Le, “A Neural Conversational Model”, vol. 37, 2016.
Review
For citations:
Nam D., Savina T. COMPUTER VISION MODEL COMPARISON. Herald of the Kazakh-British Technical University. 2021;18(3):75-82. https://doi.org/10.55452/1998-6688-2021-18-3-75-82
Views:
571
Email this article 