Otomatik inme sınıflaması: İmkansız piksellerin eliminasyonuyla etkinleştirilmiş transfer öğrenme yaklaşımı

Year 2024, , 260 - 267, 23.09.2024

İlker Özgür Koska , Çağan Koska , Antonio Fernandes

https://doi.org/10.21673/anadoluklin.1379589

Abstract

Amaç: Derin öğrenme yöntemleri ve özellikle evrişimsel sinir ağları (CNN) tıbbi görüntü sınıflamasında otomatizasyon açısından geliştirilen uygulamalarda altın standart niteliğindedir. İnme görüntülemesinde zaman oldukça kritik olup hızlı müdahale ile morbidite ve mortalite azaltılabilmektedir. Bu çalışmada amacımız hızlı inme triajı ve uygun tedavi seçimi sağlayacak iskemik inme ile hemorajik inmeyi birbirinden ayırt edebilen otomatize yöntem geliştirmektir.

Yöntemler: Teknofest sağlıkta yapay zekâ yarışması tarafından sağlanan kimliksizleştirilmiş ve anonimleştirilmiş 2000 adet iskemik inme, 2000 adet hemorajik inme içeren bilgisayarlı tomografi (BT) kesitleri kullanılarak, MobileNet ve EfficientNet CNN mimarileri transfer öğrenme metodolojisi ile, özel bir imkansız piksel değeri ve uzamsal lokalizasyonları dışlama stratejisi kullanılarak arama uzayı daraltılmış ve otomatik inme sınıflaması sağlanmıştır.

Bulgular: [0-1] normalizasyon ve 224*224’ e girişin ayarlanması dışında ön işleme yapılmayan grupta adapte MobileNetV2 ile 0.74 ve adapte EfficentNetB0 ile 0.72 doğruluk değerleri elde edildi. Öte yandan kemik yapıların çıkarıldığı ve piksel değerlerin imkânsız değerler elimine edilerek kısıtlandığı veri dönüşümü uygulanan grupta MobileNetV2 ile 0.91 ve EfficientNetB0 ile 0.88 doğruluk düzeyine ulaşıldı.

Sonuç: Derin öğrenme yöntemleri kullanılarak inme teşhisi, radyoloji uzmanı olmayan inme görüntülemeye aşina olmayan ancak inme triaj ve sağaltımında aktif rol oynayan sağlık personelleri için özellikle yararlı olabilir. Bu şekilde tedaviden fayda görecek hastanın seçimi ve tedavi kararının verilme hızı artırılabilir. Sonuç olarak iskemik-hemorajik inme sınıflandırmada yüksek doğruluk oranlarına ulaşan çalışmamız, otomatik inme tespitine katkı sağlayabilir ve hekimlerin hızlı ve uygun tedavi kararları vermelerine yardımcı olabilir.

Keywords

Derin öğrenme, inme, makine öğrenimi, yapay zekâ

Automatic stroke classification: Domain knowledge injection augmented transfer learning approach

Abstract

Aim: To build an artificial intelligence model to classify stroke into ischemic or hemorrhagic classes using the labeled stroke computer tomography (CT) slices that were shared in the 2021 Teknofest artificial intelligence in health competition.

Methods: We developed a set of methods that can inject domain knowledge into the models to provide a more refined search space for the model for better performance. We used pre-trained MobileNet and EfficientNet architectures and fine-tuned them for our 2-class output model. We discarded impossible pixel values and pixel spatial locations to provide a space that was conditioned into only possible spatial locations and signal values using our knowledge of brain anatomy, stroke pathology, and imaging.

Results: With the dataset which we just used [0-1] normalization and adjusted the input dimension into 224*224, accuracy values of 0.74 with adapted MobileNetV2 and 0.72 with adapted EfficentNetB0 were obtained in the group without further pre-processing. In the data transformation group where bone structures were removed and pixel values were restricted by eliminating impossible values, an accuracy level of 0.91 with MobileNetV2 and 0.88 with EfficientNetB0 at test time were achieved.

Conclusion: In conclusion, CT-based slice prediction of mechanism of stroke as ischemic or hemorrhagic was achieved with high accuracy by integrating human knowledge into the pre-trained off-the-shelf models which was promising to shorten the time of the triage of stroke patients which can potentially improve stroke patient outcomes.

Keywords

Artificial intelligence, stroke, machine learning, deep learning

References

• Krizhevsky A, Sutskever I, Hinton GE. ImageNet classification with deep convolutional neural networks. Communications of the ACM. 2012;60:84-90.
• Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv:1409.1556 [cs.CV]
• Szegedy, C, Liu W, Jia Y, et al. Going deeper with convolutions in: Proceedings of the IEEE conference on computer vision and pattern recognition. arXiv:1409.4842 [cs.CV]
• Kaming H, Zhang X, Residual Learning for Image Recognition. arXiv:1512.03385 [cs.CV]
• https://www.image-net.org/ Accessed date: Mar 11 2021
• https://grand-challenge.org/aiforradiology/ Radboud University Medical Center Access date: 2022
• https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection Access date: 13.11. 2019
• https://zenodo.org/record/1199398 Access date: 03.08. 2017
• Chilamkurthy S, Gosh R, Tanamala S, et al. Development and validation of deep learning algorithms for detection of critical findings in CT scans. arxiv:1803.05854 [cs.CV]
• Gulli A, Pal S (Edited by) Deep learning with Keras Packt Publishing Ltd. 2017
• Koç U, Akçapınar Sezer E, Özkaya YA, et al. Artificial intelligence in healthcare competition (TEKNOFEST-2021): Stroke data set. Eurasian J Med. 2022;54(3):248-58
• Sandler M, Howard A, Zhu M, Zhmoginov A, Chen LC. MobileNetV2: Inverted residuals and linear bottlenecks, in: 2018. IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2018;4510–20.
• Tan M, Le Q. EfficientNet: rethinking model scaling for convolutional neural networks, in: Proceedings of the 36th International Conference on Machine Learning ICML. 2019;97:6105–14.
• Nazari-Farsani S, Nyman M, Karjalainen T, Bucci M, Isojärvi J, Nummenmaa L. Automated segmentation of acute stroke lesions using a data-driven anomaly detection on diffusion weighted MRI. J Neurosci Methods. 2020;333:108575.
• Anbumozhi S. Computer aided detection and diagnosis methodology for brain stroke using adaptive neuro fuzzy inference system classifier. Int J Imaging Syst Technol. 2020;30:196–202.
• Pereira DR, Filho P, Rosa GD, Papa JP, Albuquerque VHC. Stroke lesion detection using convolutional neural networks, in: 2018 International Joint Conference on Neural Networks (IJCNN). 2018;1–6.
• Chin C, Lin B, Wu G. et al. An automated early ischemic stroke detection system using CNN deep learning algorithm, in: 2017 IEEE 8th International Conference on Awareness Science and Technology (iCAST). 2017;368-72.
• Marbun JT, Seniman U. Classification of stroke disease using convolutional neural network, J Phys: Conf Ser. 2018;978:012092
• Desai V, Flanders AE, Lakhani P. Application of deep learning in neuro- radiology: automated detection of basal ganglia hemorrhage using 2D-convolutional neural networks. arXiv:1710.03823 [cs.CV]
• Arbabshirani MR, Fornwalt BK, Mongelluzzo GJ, et al. Advanced machine learning in action: identification of intracranial hemorrhage on computed tomography scans of the head with clinical workflow integration. NPJ Digit Med. 2018;1:9
• Öman O, Mäkelä T, Salli E, Savolainen S, Kangasniemi M. 3D convolutional neural networks applied to CT angiography in the detection of acute ischemic stroke. Eur Radiol Exp. 2019;3(1):8.
• Takahashi N, Lee Y, Tsai DY, Matsuyama E, Kinoshita T, Ishii K. An automated detection method for the MCA dot sign of acute stroke in unenhanced CT. Radiol Phys Technol. 2014;7(1):79-88.
• Weber JE, Ebinger M, Rozanski M, et al. Prehospital thrombolysis in acute stroke: results of the PHANTOM-S pilot study. Neurology. 2013;80(2):163-8.