A Deep Learning-based U-Net 3+ Technique for Segmentation Blood Cell

Year 2024, Volume: 19 Issue: 2, 485 - 495, 30.09.2024

Hasan Ulutaş

https://doi.org/10.55525/tjst.1404899

Abstract

This study introduces a novel and enhanced UNet3Plus model tailored for the precise segmentation of blood cells in medical images. The architecture incorporates structural modifications, including strengthened connections between convolutional layers, increased filter numbers, and integration of Bayesian optimization for hyperparameter tuning. The model's generalization capability is optimized through the dynamic adjustment of dropout rates and learning rates. Bayesian optimization facilitates the exploration of optimal hyperparameter combinations, allowing the model to adapt effectively to diverse datasets. Advanced training strategies, such as adaptive learning rate adjustment and early stopping, are employed to mitigate overfitting and enhance training efficiency. The proposed model exhibits exceptional performance across multiple folds, achieving low training and validation losses, high accuracy metrics, and robust segmentation indices. Evaluation metrics, including Mean IoU (Jaccard Index), Dice score, Pixel Accuracy, and Precision, affirm the model's proficiency in accurately delineating blood cell boundaries. The study contributes to the field of deep learning-based medical image segmentation by showcasing the effectiveness of customized architectures and optimization techniques. The proposed UNet3Plus model stands as a promising solution for accurate and reliable blood cell segmentation, demonstrating adaptability and robust performance across various datasets. This work sets the stage for future research in the domain of medical image segmentation, emphasizing the potential for continued advancements in precise and efficient segmentation methodologies.

Keywords

blood cell, segmentation, U-Net3+, pre processing, cross validation

Kan Hücresinin Segmentasyonu için Derin Öğrenmeye Dayalı U-Net 3+ Tekniği

Abstract

Kan hücrelerinin segmentasyonu ve sınıflandırılması, hastalık teşhisi, tedavi izleme ve araştırma amaçları dahil olmak üzere çeşitli tıbbi uygulamalar için çok önemlidir. Bu süreç, farklı hücre türlerinin doğru bir şekilde tanımlanmasına ve miktarının belirlenmesine olanak tanıyarak kanla ilgili çeşitli bozuklukların tespit edilmesine ve anlaşılmasına yardımcı olur. Önerilen U-Net 3+ mimarisi, konvolüsyonel katmanlar arasındaki bağlantıları güçlendiren, filtre sayılarını artıran ve hiperparametre ayarlaması için Bayesian optimizasyonu entegre eden yapısal değişiklikler bulunmaktadır. Modelin genelleme yeteneği, dropout oranları ve öğrenme oranlarının dinamik ayarlanması ile optimize edilmiştir. Bayesian optimizasyon, optimal hiperparametre kombinasyonlarını keşfetmeyi sağlayarak modelin çeşitli veri kümelerine etkili bir şekilde uyum sağlamasına imkan tanır. Ayrıca, aşırı uydurmayı azaltmak ve eğitim verimliliğini artırmak için adaptif öğrenme oranı ayarı ve erken durdurma gibi gelişmiş eğitim stratejileri kullanılmıştır. Önerilen model, çoklu katmanlarda düşük eğitim ve doğrulama kayıpları, yüksek doğruluk metrikleri ve güçlü segmentasyon endeksleri elde ederek olağanüstü performans sergilemektedir. Değerlendirme metrikleri, ortalama IoU (Jaccard İndeksi), dice skoru, piksel doğruluğu ve hassasiyet gibi, modelin kan hücre sınırlarını doğru bir şekilde belirleme konusundaki yetkinliğini doğrular. Çalışma, özel mimarilerin ve optimizasyon tekniklerinin etkinliğini 0,9324 ortalama IoU (Jaccard İndeksi) ve 0,9667 dice skoru ile ispatlamaktadır. Önerilen U-Net 3+ modeli, çeşitli veri kümelerinde adaptasyon yeteneği ve güçlü performansıyla umut vadeden bir çözüm olarak ön plana çıkmaktadır. Bu çalışma, medikal görüntü segmentasyonu alanında gelecekteki araştırmalara zemin oluşturarak, hassas ve etkili segmentasyon metodolojilerinde devam eden ilerlemelerin potansiyelini vurgulamaktadır.

Keywords

kan hücresi, segmentasyon, UNet3+, ön işleme, çapraz doğrulama

References

• Zhu Z, et al. RETRACTED: BCNet: A Novel Network for Blood Cell Classification. Front Cell Dev Biol. 2002; (9): 813996.
• Toptaş M and Hanbay D. Mikroskobik Kan Hücre Görüntülerinin Güncel Derin Öğrenme Mimarileri ile Bölütlemesi. Mühendislik Bilimleri ve Araştırmaları Dergisi, 2023; 5(1): 135-141.
• Habibzadeh M, Jannesari M, Rezaei Z, Baharvand H, Totonchi M. Automatic white blood cell classification using pre-trained deep learning models: Resnet and inception. InTenth international conference on machine vision (ICMV 2017) 13 April 2018: SPIE. Vol. 10696, pp. 274-281.
• Sahin ME. Image processing and machine learning‐based bone fracture detection and classification using X‐ray images. Int J Imaging Syst Technol. May 2023; 33(3): 853-65.
• Şahin ME. A Deep Learning-Based Technique for Diagnosing Retinal Disease by Using Optical Coherence Tomography (OCT) Images. Turkish Journal of Science and Technology. 1 July 2022; 17(2): 417-26.
• Ulutas H, Sahin ME, Karakus MO. Application of a novel deep learning technique using CT images for COVID-19 diagnosis on embedded systems. Alexandria Eng J. 1 July 2023; 74: 345-58.
• Alam MM, Islam MT. Machine learning approach of automatic identification and counting of blood cells. Healthcare Technol Lett. August 2019; 6(4): 103-8.
• Shahin AI, Guo Y, Amin KM, Sharawi AA. White blood cells identification system based on convolutional deep neural learning networks. Comput Methods Programs Biomed. 1 January 2019; 168: 69-80.
• Banik PP, Saha R, Kim KD. An automatic nucleus segmentation and CNN model based classification method of white blood cell. Expert Syst. Appl. 1 July 2020; 149: 113211.
• Macawile MJ, Quiñones VV, Ballado A, Cruz JD, Caya MV. White blood cell classification and counting using convolutional neural network. In2018 3rd International conference on control and robotics engineering (ICCRE) 20 April 2018: IEEE. pp. 259-263.
• Yildirim M, Çinar A. Classification of white blood cells by deep learning methods for diagnosing disease. Rev d'Intelligence Artif. November 2019; 33(5): 335-40.
• Lu Y, Qin X, Fan H, Lai T, Li Z. WBC-Net: A white blood cell segmentation network based on UNet++ and ResNet. Applied Soft Computing. 1 March 2021; 101: 107006.
• Reena MR, Ameer PM. Localization and recognition of leukocytes in peripheral blood: A deep learning approach. Comput. Biol. Med. 1 November 2020; 126: 104034.
• Bozkurt F. Classification of blood cells from blood cell images using dense convolutional network. Journal of Science, Technology and Engineering Research. November 2021; 2(2): 81-8.
• Nahzat S, Bozkurt F, Yağanoğlu M. White blood cell classification using convolutional neural network. Journal of Science, Technology and Engineering Research. 2022; 3(1): 32-41.
• Khouani A, El Habib Daho M, Mahmoudi SA, Chikh MA, Benzineb B. Automated recognition of white blood cells using deep learning. Biomed Eng Lett. August 2020; 10: 359-67.
• Zhang M, Li X, Xu M, Li Q. Automated semantic segmentation of red blood cells for sickle cell disease. IEEE J Biomed Health Inf. 22 June 2020; 24(11): 3095-102.
• Dataset, https://github.com/Deponker/Blood-cell-segmentation-dataset.
• Depto DS, Rahman S, Hosen MM, Akter MS, Reme TR, Rahman A, Zunair H, Rahman MS, Mahdy MR. Automatic segmentation of blood cells from microscopic slides: a comparative analysis. Tissue and Cell. 1 December 2021; 73: 101653.
• Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation. InMedical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, 5-9 October 2015; Munich, Germany: proceedings, part III, Springer International Publishing. pp. 234-241.
• Yan X, Tang H, Sun S, Ma H, Kong D, Xie X. After-unet: Axial fusion transformer unet for medical image segmentation. InProceedings of the IEEE/CVF winter conference on applications of computer vision 2022 pp. 3971-3981.
• Huang H, Lin L, Tong R, Hu H, Zhang Q, Iwamoto Y, Han X, Chen YW, Wu J. Unet 3+: A full-scale connected unet for medical image segmentation. InICASSP 2020-2020 IEEE international conference on acoustics, speech and signal processing (ICASSP) 4 May 2020: IEEE. pp. 1055-1059.
• Deng Y, Hou Y, Yan J, Zeng D. ELU-net: An efficient and lightweight U-net for medical image segmentation. IEEE Access. 31 March 2022; 10: 35932-41.
• Sahin ME. Deep learning-based approach for detecting COVID-19 in chest X-rays. Biomed Signal Process Control. 1 September 2022; 78: 103977.
• Vasconcelos FF, Medeiros AG, Peixoto SA, Reboucas Filho PP. Automatic skin lesions segmentation based on a new morphological approach via geodesic active contour. Cognit Syst Res. 1 June 2019; 55: 44-59.
• Liu J, Yildirim O, Akin O, Tian Y. AI-driven robust kidney and renal mass segmentation and classification on 3D CT images. Bioengineering. 13 January 2023; 10(1): 116.