Branch and End Points Detection in Cerebral Vessels Images Using Deep Learning Object Detection Techniques

Yıl 2024, ERKEN GÖRÜNÜM, 1 - 1

Samet Kaya , Berna Kiraz , Ali Yılmaz Çamurcu

https://doi.org/10.2339/politeknik.1492002

Öz

In this study, we introduce a cutting-edge methodology for detecting branching and endpoints in two-dimensional brain vessel images, employing deep learning-based object detection techniques. While conventional image processing methods are viable alternatives, our adoption of deep learning showcases notable advancements in accuracy and efficiency. Following meticulous cleaning and labeling of the raw dataset sourced from laboratory environments, we meticulously convert it into the COCO format, ensuring compatibility with deep learning algorithms for both training and testing phases. Utilizing four deep learning object detection methods: fast R-CNN, faster R-CNN, RetinaNet and RPN within the Detectron2 framework, our study achieves remarkable results. Evaluation using the intersection over union (IoU) method underscores the robust performance of our deep learning approach, boasting a success rate surpassing 90%. This breakthrough not only enhances neuroimaging analysis but also holds immense potential for revolutionizing diagnostic and research practices in neurovascular studies.

Anahtar Kelimeler

Object Detection, Deep Learning, Branch and End Point Detection., Brain Vessel Analysis

Etik Beyan

This work is supported by Fatih Sultan Mehmet Vakif University Scientific Research Projects Coordination Unit under grant number 22022B1Ç01D

Destekleyen Kurum

Fatih Sultan Mehmet Vakif University

Proje Numarası

22022B1Ç01D

Teşekkür

Thanks to Fatih Sultan Mehmet Vakif University

Derin Öğrenme Nesne Tespit Teknikleri Kullanılarak Serebral Damar Görüntülerinde Dal ve Uç Noktalarının Tespiti

Öz

Bu çalışmada, derin öğrenme tabanlı nesne algılama tekniklerini kullanarak iki boyutlu beyin damarı görüntülerinde dallanma ve uç noktaları tespit etmek için son teknoloji bir metodoloji sunuyoruz. Geleneksel görüntü işleme yöntemleri uygulanabilir alternatifler olsa da, derin öğrenmeyi benimsememiz doğruluk ve verimlilikte kayda değer ilerleme sergilemekteyiz. Laboratuvar ortamlarından elde edilen ham veri setinin titizlikle temizlenmesi ve etiketlenmesinin ardından, hem eğitim hem de test aşamaları için derin öğrenme algoritmalarıyla uyumluluğu sağlamak üzere COCO formatına dönüştürüyoruz. Ardından Detectron2 çerçevesi içinde Fast R-CNN, Faster R-CNN, RetinaNet ve RPN olmak üzere dört derin öğrenme nesne algılama yöntemini kullandığımız çalışmamız dikkate değer sonuçlar elde etmektedir. Birleşme üzerinden kesişim (intersection over union) yöntemi kullanılarak yapılan değerlendirme, derin öğrenme yaklaşımımızın sağlam performansının altını çiziyor ve %90'ı aşan bir başarı oranına sahip olduğunu gösteriyoruz. Çalışma sadece nörogörüntüleme analizini geliştirmekle kalmıyor, aynı zamanda nörovasküler çalışmalarda teşhis ve araştırma uygulamalarını kolaylaştırma konusunda büyük bir potansiyele sahiptir.

Anahtar Kelimeler

Nesne Tespiti, Derin Öğrenme, Dallanma ve Uç Nokta Tespiti, Beyin Damar Analizi

Etik Beyan

Bu çalışma Fatih Sultan Mehmet Vakıf Üniversitesi Bilimsel Araştırma Projeleri Koordinatörlüğü tarafından 22022B1Ç01D hibe numarasıyla desteklenmektedir.

Destekleyen Kurum

Fatih Sultan Mehmet Vakıf Üniversitesi

Proje Numarası

22022B1Ç01D

Teşekkür

Fatih Sultan Mehmet Vakıf Üniversitesi'ne teşekkürler

Kaynakça

• [1] M. I. Todorov et al., “Automated analysis of whole brain vasculature using machine learning,” bioRxiv, pp. 0–34, (2019).
• [2] L. Y. Zhang et al., “CLARITY for high-resolution imaging and quantification of vasculature in the whole mouse brain,” Aging Dis, vol. 9, no. 2, pp. 262–272, (2018).
• [3] E. Özkan et al., “Hyperglycemia with or without insulin resistance triggers different structural changes in brain microcirculation and perivascular matrix,” Metab Brain Dis, vol. 38, no. 1, pp. 307–321, (2023).
• [4] S. Bollmann et al., “Imaging of the pial arterial vasculature of the human brain in vivo using highresolution 7T time-of-flight angiography,” Elife, vol. 11, pp. 1–35, (2022).
• [5] S. D. and A. C. and A. S. and G.-W. J. and V. I. and R. K. D. and C. Sarah. J. McGarry, “Vessel Metrics: A python based software tool for automated analysis of vascular structure in confocal imaging,” bioRxiv, vol. 151, no. 0026–2862, p. 104610, (2022).
• [6] Z. Gu et al., “CE-Net: Context Encoder Network for 2D Medical Image Segmentation,” IEEE Transactions on Medical Imaging, vol. 38, no. 10. pp. 2281–2292, (2019).
• [7] E. Zudaire, L. Gambardella, C. Kurcz, and S. Vermeren, “A computational tool for quantitative analysis of vascular networks,” PLoS One, vol. 6, no. 11, pp. 1–12, (2011).
• [8] A. Bhuiyan, B. Nath, and K. Ramamohanarao, “Detection and classification of bifurcation and branch points on retinal vascular network,” 2012 International Conference on Digital Image Computing Techniques and Applications (DICTA), pp. 1–8, (2012).
• [9] C. Anusha and P. S., “Object Detection using Deep Learning,” International Journal of Computer Applications, vol. 182, no. 32. pp. 18–22, (2018).
• [10] E. Zudaire, L. Gambardella, C. Kurcz, and S. Vermeren, “A computational tool for quantitative analysis of vascular networks,” PLoS One, vol. 6, no. 11, pp. 1–12, (2011).
• [11] F. Uslu and A. A. Bharath, “A multi-task network to detect junctions in retinal vasculature,” Lecture Notes in Computer Science, vol. 11071 LNCS, pp. 92–100, (2018).
• [12] Y. Wu, A. Kirillov, F. Massa, W.-Y. Lo, and R. Girshick, “Detectron2.” (2019).
• [13] M. I. Todorov et al., “Machine learning analysis of whole mouse brain vasculature,” Nat Methods, vol. 17, no. 4, pp. 442–449, (2020).
• [14] M. Freitas-Andrade, C. H. Comin, M. V. da Silva, L. da F. Costa, and B. Lacoste, “Unbiased analysis of mouse brain endothelial networks from two- or three-dimensional fluorescence images,” Neurophotonics, vol. 9, no. 03, (2022).
• [15] X. Ji et al., “Brain microvasculature has a common topology with local differences in geometry that match metabolic load,” Neuron, vol. 109, no. 7, pp. 1168-1187.e13, (2021).
• [16] J. Kaur and W. Singh, “A systematic review of object detection from images using deep learning,” Multimedia Tools and Applications, vol. 83, no. 4. pp. 12253–12338, (2024).
• [17] T.-Y. Lin et al., “Microsoft COCO: Common Objects in Context,” Computer Vision–ECCV 2014: 13th European Conference, vol. 8693, pp. 740–755, (2014).
• [18] Ş. Sağiroğlu and E. Beşdok, “A novel approach for image denoising based on artificial neural networks,” Journal of Polytechnic, vol. 15, no. 2. pp. 71–86, (2012).
• [19] Y. Tan, Y., Liu, M., Chen, W., Wang, X., Peng, H., & Wang, “DeepBranch: Deep Neural Networks for Branch Point Detection in Biomedical Images.” IEEE transactions on medical imaging, pp. 39(4), 1195–1205, (2020).
• [20] Z. Kuş et al., “Differential evolution-based neural architecture search for brain vessel segmentation,” Engineering Science and Technology, an International Journal, vol. 46. (2023).
• [21] N. Akyel, Cihan and Arıcı, “U-Net-RCB7: Image Segmentation Algorithm.” Journal of Polytechnic, pp. 1555–1562, (2023).
• [22] S. L. Bangare, A. Dubal, P. S. Bangare, and S. T. Patil, “Reviewing otsu’s method for image thresholding,” International Journal of Applied Engineering Research, vol. 10, no. 9, pp. 21777–21783, (2015).
• [23] Y. He, S. H. Kang, and L. Alvarez, “Finding the skeleton of 2d shape and contours: Implementation of hamilton-jacobi skeleton,” Image Processing On Line, vol. 11, no. February, pp. 18–36, (2021).
• [24] P. Murray and S. Marshall, “A Review of Recent Advances in the Hit-or-Miss Transform,” Advances in Imaging and Electron Physics, vol. 175, pp. 221–282, (2013).
• [25] S. Kaya, S. Z. Dik, B. Kiraz, M. Aydın, and A. Y. Çamurcu,“BRAINVASCULYZER: 2B Beyin Damar Görüntü Analiz Programı,” in MAS 18th International European Conference on Mathematics, Engineering, Natural & Medical Sciences, pp. 121–130, (2023).
• [26] Everingham, M., Van Gool, L., Williams, C.K.I. et al., “The PASCAL Visual Object Classes (VOC) Challenge.” Int J Comput Vis, 88, 303–338 (2010).
• [27] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), p. 779, (2016).
• [28] A. Dave, T. Khurana, P. Tokmakov, C. Schmid, and D. Ramanan, “TAO: A Large-Scale Benchmark for Tracking Any Object,” Computer Vision ECCV, pp. 436–454, (2020).
• [29] I. Scholl, T. Aach, T. M. Deserno, and T. Kuhlen, “Challenges of medical image processing,” in Computer Science-Research and Development, pp. 5–13. (2011).
• [30] J. G. Lee et al., “Deep learning in medical imaging: General overview,” Korean Journal of Radiology, vol. 18, no. 4. Korean Radiological Society, pp. 570–584, (2017).
• [31] H. P. Chan, R. K. Samala, L. M. Hadjiiski, and C. Zhou, “Deep Learning in Medical Image Analysis,” in Advances in Experimental Medicine and Biology, vol. 1213, Springer, pp. 3–2, (2020).