YOLOV9 İLE KAN HÜCRELERİNİN OTOMATİK TANIMLANMASI: OPTİMİZASYON VE ÖĞRENME ORANI ETKİLERİ

Yıl 2024, , 125 - 135, 30.04.2024

Zehra Yücel , Dilber Çetintaş

https://doi.org/10.54365/adyumbd.1388891

Öz

Kanda yer alan kan hücrelerinin mikroskobik incelenmesi zaman alıcı, pahalı ve hataya açık bir iştir. Bu çalışmanın amacı, kan hücresi görüntülerini kullanarak kan hücresi tiplerinin sınıflandırılması için YOLO mimarisini kullanan otomatik bir sistem geliştirmektir. Çalışmada kullanılan BCDD veri seti, 364 kan hücresi görüntüsü ve 4888 etiketli görüntüden oluşmaktadır. Açık kaynaklı BCCD veri seti, kırmızı kan hücrelerini (RBC'ler), beyaz kan hücrelerini (WBC'ler) ve trombositleri içerir. Geliştirilen senaryoda YOLOv9 mimarisi, farklı optimizasyon algoritmaları, öğrenme oranları kullanılarak hiperparametrelerin tanımlama sürecindeki etkisi gözlemlendi. Tanımlama sonuçlarını karşılaştırırken en iyi sonuca, 0,001 öğrenme oranıyla ADAMW optimizasyon algoritması kullanılarak ulaşıldı. Genel olarak kan hücresi tiplerinin sınıflandırılmasında WBC tanımlamasında 1,0'a yakın sonuç elde edildi. Daha sonra RBC tanımlaması yaklaşık olarak 0,93 doğrulukla elde edilirken trombositler 0,96 doğrulukla tanımlandı. Bu sonuçlar, önerilen sistemin kan hücresi tanımlamasının manuel sürecini otomatikleştirmeye yönelik etkili bir araç olarak kullanılabileceğini göstermektedir.

Anahtar Kelimeler

Kan Hücresi Tespiti, Derin Öğrenme, Yolov9, Optimizasyon, Öğrenme Oranı

AUTOMATIC IDENTIFICATION OF BLOOD CELLS WITH YOLOV9: OPTIMIZATION AND LEARNING RATE EFFECTS

Öz

Microscopic examination of blood cells in blood is time-consuming, expensive and error-prone. The aim of this study is to develop an automatic system using YOLO architecture for classification of blood cell types using blood cell images. The BCDD dataset used in the study consists of 364 blood cell images and 4888 labeled images. The open-source BCCD dataset includes red blood cells (RBCs), white blood cells (WBCs), and platelets. In the developed scenario, the YOLO v9 architecture, different optimization algorithms, and learning rates were used to observe the effect of hyperparameters in the parameter definition process. When comparing the identification results, the best result was achieved using the ADAMW optimization algorithm with a learning rate of 0.001. Overall, a result close to 1.0 was obtained in the WBC classification for blood cell types. Subsequently, the RBC identification was achieved with an accuracy of approximately 0.93, while platelets were identified with an accuracy of 0.96. These results indicate that the proposed system could be used as an effective tool for automating the manual process of blood cell identification.

Anahtar Kelimeler

Blood Cell Detection, Deep Learning, Yolov9, Optimization, Learning Rate

Kaynakça

• Alberts, B. (2017). Molecular biology of the cell. Garland science.
• Alomari, Y. M., Sheikh Abdullah, S. N. H., Zaharatul Azma, R., & Omar, K. (2014). Automatic detection and quantification of WBCs and RBCs using iterative structured circle detection algorithm. Computational and mathematical methods in medicine, 2014.
• Faggio C, Sureda A, Morabito S, Sanches-Silva A, Mocan A, Nabavi SF, Nabavi SM. Flavonoids and platelet aggregation: A brief review. European journal of pharmacology. 2017 Jul 15;807:91-101.
• Farag MR, Alagawany M. Erythrocytes as a biological model for screening of xenobiotics toxicity. Chemico-biological interactions. 2018 Jan 5;279:73-83.
• Rezatofighi SH, Soltanian-Zadeh H. Automatic recognition of five types of white blood cells in peripheral blood. Computerized Medical Imaging and Graphics. 2011 Jun 1;35(4):333-43.
• Acharjee, S., Chakrabartty, S., Alam, M. I., Dey, N., Santhi, V., & Ashour, A. S. (2016, March). A semiautomated approach using GUI for the detection of red blood cells. In 2016 International conference on electrical, electronics, and optimization techniques (ICEEOT) (pp. 525-529). IEEE.
• Arslan, Özkan, and Mustafa Karhan. "Effect of Hilbert-Huang transform on classification of PCG signals using machine learning." Journal of King Saud University-Computer and Information Sciences (2022).
• Yaman, O., & Tuncer, T. (2022). Exemplar pyramid deep feature extraction based cervical cancer image classification model using pap-smear images. Biomedical Signal Processing and Control, 73, 103428.
• Habibzadeh, Mehdi, Adam Krzyżak, and Thomas Fevens. "White blood cell differential counts using convolutional neural networks for low resolution images." International Conference on Artificial Intelligence and Soft Computing. Springer, Berlin, Heidelberg, 2013.
• Vatathanavaro, Supawit, Suchat Tungjitnob, and Kitsuchart Pasupa. "White blood cell classification: a comparison between VGG-16 and ResNet-50 models." proceeding of the 6th joint symposium on computational intelligence (JSCI6). Vol. 12. 2018.
• Diouf, Daouda, et al. "Convolutional Neural Network and decision support in medical imaging: case study of the recognition of blood cell subtypes." arXiv preprint arXiv:1911.08010 (2019).
• Zhao, Jianwei, et al. "Automatic detection and classification of leukocytes using convolutional neural networks." Medical & biological engineering & computing 55.8 (2017): 1287-1301.
• Özyurt, Fatih. "A fused CNN model for WBC detection with MRMR feature selection and extreme learning machine." Soft Computing 24.11 (2020): 8163-8172.
• BCCD Veri seti. https://github.com/Shenggan/BCCD_Dataset (Erişim tarihi: Şubat 2024).
• Redmon, Joseph, et al. "You only look once: Unified, real-time object detection." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
• Lin, Tsung-Yi, et al. "Microsoft coco: Common objects in context." European conference on computer vision. Springer, Cham, 2014.
• Shinde, Shubham, Ashwin Kothari, and Vikram Gupta. "YOLO based human action recognition and localization." Procedia computer science 133 (2018): 831-838.
• Wang CY, Yeh IH, Liao HY. YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information. arXiv preprint arXiv:2402.13616. 2024 Feb 21.
• YOLOv9. https://docs.ultralytics.com/tr/models/yolov9/ (Erişim Tarihi, Mart,2024).
• Liu S, Qi L, Qin H, Shi J, Jia J. Path aggregation network for instance segmentation. InProceedings of the IEEE conference on computer vision and pattern recognition 2018 (pp. 8759-8768).
• Cai Y, Zhou Y, Han Q, Sun J, Kong X, Li J, Zhang X. Reversible column networks. arXiv preprint arXiv:2212.11696. 2022 Dec 22.
• Lin, Tsung-Yi, et al. "Microsoft coco: Common objects in context." European conference on computer vision. Springer, Cham, 2014.
• Bottou L. Large-scale machine learning with stochastic gradient descent. InProceedings of COMPSTAT'2010: 19th International Conference on Computational StatisticsParis France, August 22-27, 2010 Keynote, Invited and Contributed Papers 2010 (pp. 177-186). Physica-Verlag HD.
• Kingma DP, Ba J. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980. 2014 Dec 22.
• Loshchilov I, Hutter F. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101. 2017 Nov 14.