نشریه علمی مهندسی پزشکی زیستی
Iranian Journal of Biomedical Engineering (IJBME)

جست‌وجوی معماری شبکه‌های عصبی UNet با استفاده از الگوریتم ژنتیک به منظور استخراج تومور سرطان سینه در تصاویر اولتراسوند (GNAS-UNet)

نوع مقاله : مقاله کامل پژوهشی

نویسندگان

شیما منصوب‌ریحانیان 1
مه‌لقا افراسیابی 2
علی دلشادی 1
سمیرا عباسی 3

1 کارشناسی ارشد، گروه مهندسی پزشکی، دانشگاه صنعتی همدان، همدان، ایران

2 استادیار، گروه کامپیوتر، دانشکده‌ی برق و کامپیوتر، دانشگاه صنعتی همدان، همدان، ایران

3 استادیار، گروه مهندسی پزشکی، دانشگاه صنعتی همدان، همدان، ایران

https://doi.org/10.22041/ijbme.2025.2032100.1907
چکیده
تقسیم­‌بندی دقیق بافت سینه در تصاویر اولتراسوند برای تشخیص زودهنگام و تحلیل سرطان سینه بسیار حائز اهمیت است. روش‌­های سنتی برای تقسیم­‌بندی تصاویر پزشکی غالبا به طراحی دستی معماری‌­های شبکه‌ی عصبی وابسته هستند، که می‌­تواند زمان‌بر و متکی به دانش کارشناسان باشد. در این مطالعه رویکرد جدیدی پیشنهاد شده است که در آن جست‌وجوی معماری شبکه (NAS) با الگوریتم ژنتیک ترکیب شده تا شبکه‌ی عصبی عمیق UNet برای تقسیم­‌بندی بافت سینه بهینه‌‌سازی شود. این ترکیب از مزایای هر دو جست‌وجوی معماری شبکه و الگوریتم‌­های تکاملی بهره‌برداری می­کند، تا عمل‌کرد مدل را بهبود بخشد و امکان طراحی خودکار معماری کارآمد را فراهم سازد. نتایج تجربی نشان می‌­دهد که رویکرد پیشنهادی در بهینه‌­سازی فراپارامتر­ها موثر و کارآمد است. مدل‌­های UNet که به صورت خودکار ساخته شده­‌اند، از لحاظ دقت تقسیم­‌بندی و کارایی محاسباتی با معماری‌­های طراحی شده‌ی دستی قابل رقابت هستند. عمل‌کرد روش پیشنهادی (GNAS-UNet) روی مجموعه‌ی داده‌ی BUSI ارزیابی شده است. نمرات DC و Miou به ترتیب برابر با 937/0 و 906/0 به دست آمده که نشان ‌دهنده‌ی رقابت­‌پذیری معماری‌های خودکار با معماری‌­های دست­‌ساز است. علاوه بر این، توانایی الگوریتم ژنتیک در کاوش فضای جست‌وجو به شناسایی پیکر­بندی­‌های شبکه­‌ای مناسب که به خوبی به داده‌­های جدید تعمیم می‌­یابند منجر شده است. نتایج این پژوهش بر پتانسیل تکنیک­‌های یادگیری ماشین خودکار در پیش‌برد دقت و کارایی وظایف تقسیم­‌بندی تصاویر پزشکی تاکید دارد. یکی از اهداف اصلی این تحقیق کاهش زمان طراحی و ارائه‌ی رویکردی کارآمد برای ساخت معماری‌های بهینه است.

کلیدواژه‌ها

تقسیم‌بندی تصاویر
معماری عصبی
الگوریتم ژنتیک
تصاویر اولتراسوند
تومور سینه

موضوعات

عنوان مقاله English

Neural Architecture Search based on UNet using Genetic Algorithm for Extraction of Breast Cancer in Ultrasound Images (GNAS-UNet)

نویسندگان English

Shima MansoubReyhanian 1
Mahlagha Afrasiabi 2
Ali Delshadi, 1
Samira Abbasi 3
1 M.Sc., Biomedical Engineering Department, Hamedan University of Technology, Hamedan, Iran
2 Assistant Professor, Computer Engineering, Hamedan University of Technology, Hamedan, Iran
3 Assistant Professor, Biomedical Engineering Department, Hamedan University of Technology, Hamedan, Iran
چکیده English

Accurate segmentation of breast tissue in ultrasound images is crucial for the early detection and analysis of breast cancer. Traditional approaches to medical image segmentation often rely on manually designed neural network architectures, which can be time-consuming and dependent on expert knowledge. In this study, we propose a novel approach that combines Neural Architecture Search (NAS) with a genetic algorithm to optimize the UNet deep neural network for breast tissue segmentation. This combination leverages the strengths of both NAS and evolutionary algorithms to enhance model performance and enable the automated design of efficient architectures. Experimental results demonstrate that our proposed approach is both effective and efficient in optimizing hyperparameters. The automatically constructed UNet models are competitive with manually designed architectures in terms of segmentation accuracy and computational efficiency. The GNAS-UNet model is utilized on the Breast Ultrasound Image Dataset (BUSI). The experimental findings demonstrate competitive performance, achieving an overall DC of 0.937, Miou of 0.906, demonstrating the competitiveness of automated architectures with manually crafted ones. Furthermore, the genetic algorithm's ability to navigate the search space results in the discovery of suitable network configurations that generalize well to new data. Our results underscore the potential of automated machine learning techniques in advancing the accuracy and efficiency of medical image segmentation tasks. One of the primary objectives of this study is to reduce the design time and provide an efficient approach for building optimal architectures.

کلیدواژه‌ها English

Image Segmentation
Neural Architecture
Genetic Algorithm
Ultrasound Images
Breast Tumor
  1. Jin, L. Cruz, and N. Gonçalves, “Deep facial diagnosis: deep transfer learning from face recognition to facial diagnosis,” IEEE Access, vol. 8, pp. 123649–123661, 2020.
  2. Yao et al., “Compound figure separation of biomedical images with side loss,” in Deep Generative Models, and Data Augmentation, Labelling, and Imperfections: First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Strasbourg, France, October 1, 2021, Proceedings 1, 2021, pp. 173–183.
  3. Zhao et al., “VoxelEmbed: 3D instance segmentation and tracking with voxel embedding based deep learning,” in Machine Learning in Medical Imaging: 12th International Workshop, MLMI 2021, Held in Conjunction with MICCAI 2021, Strasbourg, France, September 27, 2021, Proceedings 12, 2021, pp. 437–446.
  4. Minaee, Y. Boykov, F. Porikli, A. Plaza, N. Kehtarnavaz, and D. Terzopoulos, “Image segmentation using deep learning: A survey,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 44, no. 7, pp. 3523–3542, 2021.
  5. Rahman, J. M. J. Valanarasu, I. Hacihaliloglu, and V. M. Patel, “Ambiguous medical image segmentation using diffusion models,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 11536–11546.
  6. Kratkiewicz, A. Pattyn, N. Alijabbari, and M. Mehrmohammadi, “Ultrasound and photoacoustic imaging of breast cancer: clinical systems, challenges, and future outlook,” J. Clin. Med., vol. 11, no. 5, p. 1165, 2022.
  7. Suganyadevi, V. Seethalakshmi, and K. Balasamy, “A review on deep learning in medical image analysis,” Int. J. Multimed. Inf. Retr., vol. 11, no. 1, pp. 19–38, 2022.
  8. Sun, B. Xue, M. Zhang, G. G. Yen, and J. Lv, “Automatically designing CNN architectures using the genetic algorithm for image classification,” IEEE Trans. Cybern., vol. 50, no. 9, pp. 3840–3854, 2020.
  9. Rajakumari and L. Kalaivani, “Breast Cancer Detection and Classification Using Deep CNN Techniques.,” Intell. Autom. Soft Comput., vol. 32, no. 2, 2022.
  10. Yadavendra and S. Chand, “A comparative study of breast cancer tumor classification by classical machine learning methods and deep learning method,” Mach. Vis. Appl., vol. 31, no. 6, p. 46, 2020.
  11. Zhang, L. Han, K. Chen, Y. Peng, and J. Lin, “Diagnostic efficiency of the breast ultrasound computer-aided prediction model based on convolutional neural network in breast cancer,” J. Digit. Imaging, vol. 33, pp. 1218–1223, 2020.
  12. M. de Miranda Almeida, D. Chen, A. L. da Silva Filho, and W. C. Brandao, “Machine Learning Algorithms for Breast Cancer Detection in Mammography Images: A Comparative Study.,” in ICEIS (1), 2021, pp. 660–667.
  13. Zoph and Q. V Le, “Neural architecture search with reinforcement learning,” arXiv Prepr. arXiv1611.01578, 2016.
  14. Jaafra, J. L. Laurent, A. Deruyver, and M. S. Naceur, “Reinforcement learning for neural architecture search: A review,” Image Vis. Comput., vol. 89, pp. 57–66, 2019.
  15. Cai, L. Chen, S. Zeng, Y. Lai, and H. Liu, “EST-NAS: An evolutionary strategy with gradient descent for neural architecture search,” Appl. Soft Comput., vol. 146, p. 110624, 2023.
  16. Lv, X. Song, Y. Feng, Y. Ou, Y. Sun, and M. Zhang, “Evolutionary Neural Network Architecture Search,” in Handbook of Evolutionary Machine Learning, Springer, 2023, pp. 247–281.
  17. Ren et al., “Convolutional neural network detection of axillary lymph node metastasis using standard clinical breast MRI,” Clin. Breast Cancer, vol. 20, no. 3, pp. e301–e308, 2020.
  18. Raza, N. Ullah, J. A. Khan, M. Assam, A. Guzzo, and H. Aljuaid, “DeepBreastCancerNet: A novel deep learning model for breast cancer detection using ultrasound images,” Appl. Sci., vol. 13, no. 4, p. 2082, 2023.
  19. Liu, L. Song, S. Liu, and Y. Zhang, “A review of deep-learning-based medical image segmentation methods,” Sustainability, vol. 13, no. 3, p. 1224, 2021.
  20. Chitnis, R. Hosseini, and P. Xie, “Brain tumor classification based on neural architecture search,” Sci. Rep., vol. 12, no. 1, p. 19206, 2022.
  21. Yang et al., “A 3D prediction model for benign or malignant of pulmonary nodules based on neural architecture search,” Signal, Image Video Process., vol. 18, no. 1, pp. 843–852, 2024.
  22. -S. Kang, J. Kang, J.-J. Kim, K.-W. Jeon, H.-J. Chung, and B.-H. Park, “Neural Architecture Search Survey: A Computer Vision Perspective,” Sensors, vol. 23, no. 3. 2023. doi: 10.3390/s23031713.
  23. Wang and W. Zhu, “Advances in neural architecture search,” Natl. Sci. Rev., vol. 11, no. 8, p. nwae282, Aug. 2024, doi: 10.1093/nsr/nwae282.
  24. Črepinšek, S. Liu, and M. Mernik, “Exploration and Exploitation in Evolutionary Algorithms: A Survey,” ACM Comput. Surv., vol. 45, p. Article 35, Jun. 2013, doi: 10.1145/2480741.2480752.
  25. Real et al., “Large-scale evolution of image classifiers,” in International conference on machine learning, 2017, pp. 2902–2911.
  26. Chen, T. Yang, X. Zhang, G. Meng, X. Xiao, and J. Sun, “Detnas: Backbone search for object detection,” Adv. Neural Inf. Process. Syst., vol. 32, 2019.
  27. Li et al., “A generalized framework for population based training,” in Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2019, pp. 1791–1799.
  28. R. Young, D. C. Rose, T. P. Karnowski, S.-H. Lim, and R. M. Patton, “Optimizing deep learning hyper-parameters through an evolutionary algorithm,” in Proceedings of the workshop on machine learning in high-performance computing environments, 2015, pp. 1–5.
  29. U. Nasir, S. Earle, J. Togelius, S. James, and C. Cleghorn, “LLMatic: neural architecture search via large language models and quality diversity optimization,” in Proceedings of the Genetic and Evolutionary Computation Conference, 2024, pp. 1110–1118.
  30. Han, Y. Xue, Z. Wang, Y. Zhang, A. Muravev, and M. Gabbouj, “SaDENAS: A self-adaptive differential evolution algorithm for neural architecture search,” Swarm Evol. Comput., vol. 91, p. 101736, 2024.
  31. Shao, “A Survey on Neural Architecture Search Based on Reinforcement Learning,” arXiv Prepr. arXiv2409.18163, 2024.
  32. S. Carras, C. Pereira, D. Biswas, C. Lee, S. Partridge, and A. M. Alessio, “Genetic algorithm for machine learning architecture selection for breast MRI classification,” in Medical Imaging 2020: Computer-Aided Diagnosis, 2020, vol. 11314, pp. 488–493.
  33. N. Oyelade and A. E. Ezugwu, “A bioinspired neural architecture search based convolutional neural network for breast cancer detection using histopathology images,” Sci. Rep., vol. 11, no. 1, p. 19940, 2021.
  34. Zhang et al., “Fully automatic tumor segmentation of breast ultrasound images with deep learning,” J. Appl. Clin. Med. Phys., vol. 24, Dec. 2022, doi: 10.1002/acm2.13863.
  35. Liu, J. Shi, and Q. Zhang, “Tumor classification by deep polynomial network and multiple kernel learning on small ultrasound image dataset,” Lect. Notes Comput. Sci. (including Subser. Lect. Notes Artif. Intell. Lect. Notes Bioinformatics), vol. 9352, pp. 313–320, 2015, doi: 10.1007/978-3-319-24888-2_38.
  36. Liu, S. Xie, J. Yu, L. Niu, and W. Sun, “Classification of thyroid nodules in ultrasound images using deep model based transfer learning and hybrid features,” in 2017 IEEE international conference on acoustics, speech and signal processing (ICASSP), 2017, pp. 919–923.
  37. Han et al., “A deep learning framework for supporting the classification of breast lesions in ultrasound images,” Phys. Med. Biol., vol. 62, no. 19, p. 7714, 2017.
  38. Huang et al., “Unet 3+: A full-scale connected unet for medical image segmentation,” in ICASSP 2020-2020 IEEE international conference on acoustics, speech and signal processing (ICASSP), 2020, pp. 1055–1059.
  39. Weng, T. Zhou, Y. Li, and X. Qiu, “Nas-unet: Neural architecture search for medical image segmentation,” IEEE access, vol. 7, pp. 44247–44257, 2019.
  40. Wu and K. He, “Group normalization,” in Proceedings of the European conference on computer vision (ECCV), 2018, pp. 3–19.
  41. M. Shareef et al., “A Benchmark for Breast Ultrasound Image Classification,” Available SSRN 4339660, 2023.
  42. S. Punn and S. Agarwal, “RCA-IUnet: a residual cross-spatial attention-guided inception U-Net model for tumor segmentation in breast ultrasound imaging,” Mach. Vis. Appl., vol. 33, no. 2, p. 27, 2022.
نشریه علمی مهندسی پزشکی زیستی
دوره 18، شماره 1
بهار 1403
صفحه 35-49

  • تاریخ دریافت 23 خرداد 1403
  • تاریخ بازنگری 19 آذر 1403
  • تاریخ پذیرش 29 دی 1403

صفحه اصلی

تماس با ما