.: ORİJİNAL ARAŞTIRMA
Evaluation of Computed Tomography Images by Convolutional Neural Network Architectures for COVID-19 Diagnosis: A Methodological Study
COVID-19 Tanısında Bilgisayarlı Tomografi Görüntülerinin Konvolüsyonel Sinir Ağları Mimarileri ile Değerlendirilmesi: Metodolojik Çalışma
Turkiye Klinikleri J Biostat. 2025;17(1):16-27
doi: 10.5336/biostatic.2024-106714
Makale Dili: EN
Tam Metin
ABSTRACT
Objective: For the diagnosis of coronavirus disease2019 (COVID-19) using computed tomography (CT) images, disease classification from images must be performed with high accuracy. This study aims to evaluate the classification success on CT images using fine-tuned convolutional neural network (CNN) architectures applied to the new dataset with transfer learning method, which provides accurate and fast diagnosis of COVID-19, by comparing their classification success with various performance measures. Material and Methods: The classification of the disease was performed with CNN architectures Xception, InceptionResNetV2, InceptionV3, ResNet50, VGG16, VGG19, MobileNetV2, DenseNet121, DenseNet169 and DenseNet201. The performance of the architectures was utilized with accuracy, sensitivity, specificity, precision, F1 score and Area Under the Curve (AUC). In total, the data included 7,593 CT images of 466 COVID-19 patients and 6,893 CT images of 604 nonCOVID-19 patients. Results: The architecture that had the highest performance was InceptionResNetV2, with an accuracy of 98.00%, sensitivity of 98.53%, specificity of 97.53%, precision of 97.25%, F1 score of 97.88%, and AUC of 98.03%. ResNet50 showed the lowest performance with accuracy of 97.14%, sensitivity of 98.75%, specificity of 95.70%, precision of 95.32%, F1 score of 97.01% and AUC of 97.23%. Conclusion: All architectures examined were found to work with performance measures above 95% in COVID-19 diagnostics. As a result, it was found that the preprocessing, fine-tuning, and hyperparameter optimization of the architectures we used are generalizable for COVID-19 and chest CT images, and that each of the architectures works with good performance.
Keywords: Coronavirus disease-2019; computed tomography; image classification; deep learning; convolutional neural networks
Objective: For the diagnosis of coronavirus disease2019 (COVID-19) using computed tomography (CT) images, disease classification from images must be performed with high accuracy. This study aims to evaluate the classification success on CT images using fine-tuned convolutional neural network (CNN) architectures applied to the new dataset with transfer learning method, which provides accurate and fast diagnosis of COVID-19, by comparing their classification success with various performance measures. Material and Methods: The classification of the disease was performed with CNN architectures Xception, InceptionResNetV2, InceptionV3, ResNet50, VGG16, VGG19, MobileNetV2, DenseNet121, DenseNet169 and DenseNet201. The performance of the architectures was utilized with accuracy, sensitivity, specificity, precision, F1 score and Area Under the Curve (AUC). In total, the data included 7,593 CT images of 466 COVID-19 patients and 6,893 CT images of 604 nonCOVID-19 patients. Results: The architecture that had the highest performance was InceptionResNetV2, with an accuracy of 98.00%, sensitivity of 98.53%, specificity of 97.53%, precision of 97.25%, F1 score of 97.88%, and AUC of 98.03%. ResNet50 showed the lowest performance with accuracy of 97.14%, sensitivity of 98.75%, specificity of 95.70%, precision of 95.32%, F1 score of 97.01% and AUC of 97.23%. Conclusion: All architectures examined were found to work with performance measures above 95% in COVID-19 diagnostics. As a result, it was found that the preprocessing, fine-tuning, and hyperparameter optimization of the architectures we used are generalizable for COVID-19 and chest CT images, and that each of the architectures works with good performance.
Keywords: Coronavirus disease-2019; computed tomography; image classification; deep learning; convolutional neural networks
ÖZET
Amaç: Bilgisayarlı tomografi (BT) görüntüleri kullanılarak koronavirüs hastalığı [coronavirus disease-2019 (COVID-19)] tanısı için görüntülerden hastalık sınıflandırmasının yüksek doğrulukla yapılması gerekmektedir. Bu çalışma, COVID-19'un doğru ve hızlı tanısını sağlayan, transfer öğrenme yöntemi ile yeni veri setine uygulanan ince ayarlı konvolüsyonel sinir ağları (KSA) mimarilerini kullanarak BT görüntüleri üzerinde sınıflandırma başarılarını çeşitli performans ölçütleri ile karşılaştırılarak değerlendirmeyi amaçlamaktadır. Gereç ve Yöntemler: Hastalığın sınıflandırılması KSA mimarilerinden Xception, InceptionResNetV2, InceptionV3, ResNet50, VGG16, VGG19, MobileNetV2, DenseNet121, DenseNet169 ve DenseNet201 kullanılarak gerçekleştirildi. Mimarilerin performansları doğruluk, duyarlık, seçicilik, kesinlik, F1 skoru ve Eğrinin Altındaki Alan [Area Under the Curve (AUC)] ile değerlendirildi. Kullanılan veri, 466 COVID-19 hastasının 7.593 BT görüntüsünü ve 604 COVID-19 olmayan hastanın 6.893 BT görüntüsünü içeriyordu. Bulgular: En yüksek performansa sahip mimari, %98,00 doğruluk, %98,53 duyarlık, %97,53 seçicilik, 97,25 kesinlik, 97,88 F1 puanı ve 98,03 AUC ile InceptionResNetV2 oldu. En düşük performansı %97,14 doğruluk, %98,75 duyarlık, %95,70 seçicilik, %95,32 kesinlik, %97,01 F1 skoru ve %97,23 AUC ile ResNet50 gösterdi. Sonuç: İncelenen tüm mimarilerin %95'in üzerinde performans ölçüleriyle çalıştığı görüldü. Sonuç olarak, kullandığımız mimarilerin ön işleme, ince ayar ve hiperparametre optimizasyonunun COVID-19 ve göğüs BT görüntüleri için genelleştirilebilir olduğu ve mimarilerin her birinin iyi performansla çalıştığı görülmüştür.
Anahtar Kelimeler: Koronavirüs hastalığı-2019; bilgisayarlı tomografi; görüntü sınıflandırma; derin öğrenme; konvolüsyonel sinir ağları
Amaç: Bilgisayarlı tomografi (BT) görüntüleri kullanılarak koronavirüs hastalığı [coronavirus disease-2019 (COVID-19)] tanısı için görüntülerden hastalık sınıflandırmasının yüksek doğrulukla yapılması gerekmektedir. Bu çalışma, COVID-19'un doğru ve hızlı tanısını sağlayan, transfer öğrenme yöntemi ile yeni veri setine uygulanan ince ayarlı konvolüsyonel sinir ağları (KSA) mimarilerini kullanarak BT görüntüleri üzerinde sınıflandırma başarılarını çeşitli performans ölçütleri ile karşılaştırılarak değerlendirmeyi amaçlamaktadır. Gereç ve Yöntemler: Hastalığın sınıflandırılması KSA mimarilerinden Xception, InceptionResNetV2, InceptionV3, ResNet50, VGG16, VGG19, MobileNetV2, DenseNet121, DenseNet169 ve DenseNet201 kullanılarak gerçekleştirildi. Mimarilerin performansları doğruluk, duyarlık, seçicilik, kesinlik, F1 skoru ve Eğrinin Altındaki Alan [Area Under the Curve (AUC)] ile değerlendirildi. Kullanılan veri, 466 COVID-19 hastasının 7.593 BT görüntüsünü ve 604 COVID-19 olmayan hastanın 6.893 BT görüntüsünü içeriyordu. Bulgular: En yüksek performansa sahip mimari, %98,00 doğruluk, %98,53 duyarlık, %97,53 seçicilik, 97,25 kesinlik, 97,88 F1 puanı ve 98,03 AUC ile InceptionResNetV2 oldu. En düşük performansı %97,14 doğruluk, %98,75 duyarlık, %95,70 seçicilik, %95,32 kesinlik, %97,01 F1 skoru ve %97,23 AUC ile ResNet50 gösterdi. Sonuç: İncelenen tüm mimarilerin %95'in üzerinde performans ölçüleriyle çalıştığı görüldü. Sonuç olarak, kullandığımız mimarilerin ön işleme, ince ayar ve hiperparametre optimizasyonunun COVID-19 ve göğüs BT görüntüleri için genelleştirilebilir olduğu ve mimarilerin her birinin iyi performansla çalıştığı görülmüştür.
Anahtar Kelimeler: Koronavirüs hastalığı-2019; bilgisayarlı tomografi; görüntü sınıflandırma; derin öğrenme; konvolüsyonel sinir ağları
REFERANSLAR:
- Gunraj H, Sabri A, Koff D, Wong A. COVID-Net CT-2: enhanced deep neural networks for detection of COVID-19 from chest CT images through bigger, more diverse learning. Front Med (Lausanne). 2022;8:729287. [Crossref] [PubMed] [PMC]
- Demir FB, Yılmaz E. X-ray görüntülerinden COVID-19 tespiti için derin öğrenme temelli bir yaklaşım [A deep learning-based approach for detection of COVID-19 from X-ray images]. Avrupa Bilim ve Teknoloji Dergisi. 2021;(32):627-32. [Crossref]
- Ai T, Yang Z, Hou H, Zhan C, Chen C, Lv W, et al. Correlation of chest CT and RT-PCR testing for coronavirus disease 2019 (COVID-19) in China: a report of 1014 cases. Radiology. 2020;296(2):E32-E40. [Crossref] [PubMed] [PMC]
- Dong D, Tang Z, Wang S, Hui H, Gong L, Lu Y, et al. The role of imaging in the detection and management of COVID-19: a review. IEEE Rev Biomed Eng. 2021;14:16-29. [Crossref] [PubMed]
- Islam MM, Karray F, Alhajj R, Zeng J. A review on deep learning techniques for the diagnosis of novel coronavirus (COVID-19). IEEE Access. 2021;9:30551-72. [Crossref] [PubMed] [PMC]
- Clark S, Kamalinejad E, Magpantay C, Sahota S, Zhong J, Hu Y. A review of CNN on medical imaging to diagnose COVID-19 infections. 34th International Conference on Computer Applications in Industry and Engineering. CAINE. 2021;79:91-8. [Link]
- Serte S, Demirel H. Deep learning for diagnosis of COVID-19 using 3D CT scans. Comput Biol Med. 2021;132:104306. [Crossref] [PubMed] [PMC]
- Kugunavar S, Prabhakar CJ. Convolutional neural networks for the diagnosis and prognosis of the coronavirus disease pandemic. Vis Comput Ind Biomed Art. 2021;4(1):12. [Crossref] [PubMed] [PMC]
- Riyadi S, Septiari TD, Damarjati C, Ramli S. Covid-19 detection based-on CT-scan images using inception deep learning. 2022 IEEE 12th International Conference on Control System, Computing and Engineering (ICCSCE) Penang, Malaysia. 2022. p.214-8. [Crossref]
- Maftouni M, Law ACC, Shen B, Kong Grado Z, Zhou Y, Yazdi NA. A robust ensemble-deep learning model for COVID-19 diagnosis based on an integrated CT scan images database. In Ghate A, Krishnaiyer K, Paynabar K, eds. IISE Annual Conference and Expo 2021. Institute of Industrial and Systems Engineers, IISE. 2021. p.632-7.
- Taye MM. Theoretical understanding of convolutional neural network: concepts, architectures, applications, future directions. Computation. 2023;11(3):52. [Crossref]
- Medium [İnternet]. Convolutional Neural Network (ConvNet yada CNN) nedir, nasıl çalışır? 2020. [Cited: September 26, 2024]. Available from: [Link]
- Yamashita R, Nishio M, Do RKG, Togashi K. Convolutional neural networks: an overview and application in radiology. Insights Imaging. 2018;9(4):611-29. [Crossref] [PubMed] [PMC]
- Chollet F. Xception: deep learning with depthwise separable convolutions. CVPR. IEEE Computer Society. 2017. p.1800-7. [Link]
- Varaich ZA, Khalid S. Recognizing actions of distracted drivers using inception v3 and Xception convolutional neural networks. IEEE. 2019. pp.1-8. [Crossref]
- Wani MA, Bhat FA, Afzal S, Khan AI. Basics of Supervised Deep Learning. Advances in Deep Learning. Berlin: Springer; 2020. vol. 57. p.13-29. [Crossref]
- Shazia A, Xuan TZ, Chuah JH, Usman J, Qian P, Lai KW. A comparative study of multiple neural network for detection of COVID-19 on chest X-ray. EURASIP J Adv Signal Process. 2021;2021(1):50. [Crossref] [PubMed] [PMC]
- Jaiswal A, Gianchandani N, Singh D, Kumar V, Kaur M. Classification of the COVID-19 infected patients using DenseNet201 based deep transfer learning. J Biomol Struct Dyn. 2021;39(15):5682-9. [Crossref] [PubMed]
- Khattab R, Abdelmaksoud IR, Abdelrazek S. Automated detection of COVID-19 and pneumonia diseases using data mining and transfer learning algorithms with focal loss from chest X-ray images. Applied Soft Computing. 2024;162:111806-6. [Crossref]
- Nguyen LD, Lin D, Lin Z, Cao J. Deep CNNs for microscopic image classification by exploiting transfer learning and feature concatenation. 2018 IEEE International Symposium on Circuits and Systems (ISCAS). 2018. pp.1-5. [Crossref]
- Issahaku FY, Liu X, Lu K, Fang X, Danwana SB, Asimeng E. Multimodal deep learning model for COVID-19 detection. Biomedical Signal Processing and Control. 2024;91:105906-6. [Crossref]
- Toğaçar M, Ergen B, Cömert Z. Application of breast cancer diagnosis based on a combination of convolutional neural networks, ridge regression and linear discriminant analysis using invasive breast cancer images processed with autoencoders. Med Hypotheses. 2020;135:109503. [Crossref] [PubMed]
- Theckedath D, Sedamkar RR. Detecting affect states using VGG16, ResNet50 and SE-ResNet50 networks. SN Comput Sci. 2020;1(2). [Crossref]
- Asif S, Qurrat-ul-Ain, Awais M, Amjad K, Bilal O, Al-Sabri R, et al. A deep ensemble learning framework for COVID-19 detection in chest X-ray images. Netw Model Anal Health Inform Bioinform. 2024;13(1):30. [Crossref]
- Matsuyama E. A deep learning interpretable model for novel coronavirus disease (COVID-19) screening with chest CT images. J. Biomed.Sci. Eng. 2020;13(7):140-52. [Crossref]
- He X, Yang X, Zhang S, Zhao J, Zhang Y, Xing E, et al. Sample-efficient deep learning for COVID-19 diagnosis based on CT scans. Medrxiv 2020:1-9. [Crossref]
- Yang S, Jiang L, Cao Z, Wang L, Cao J, Feng R, et al. Deep learning for detecting corona virus disease 2019 (COVID-19) on high-resolution computed tomography: a pilot study. Ann Transl Med. 2020;8(7):450. [Crossref] [PubMed] [PMC]
- Mishra NK, Singh P, Joshi SD. Automated detection of COVID-19 from CT scan using convolutional neural network. Biocybern Biomed Eng. 2021;41(2):572-88. [Crossref] [PubMed] [PMC]
- Shaik NS, Cherukuri TK. Transfer learning based novel ensemble classifier for COVID-19 detection from chest CT-scans. Comput Biol Med. 2022;141:105127. [Crossref] [PubMed] [PMC]
- Bouzid HAI, Berrani SA, Yahiaoui S, Belaid A, Belazzougui D, Djouad M, et al. Deep learning-based COVID-19 diagnosis: a thorough assessment with a focus on generalization capabilities. J Image Video Proc. 2024;40. [Crossref]
- Appavu N, Babu CNK, Kadry S. COVID-19 classification in X-ray/CT images using pretrained deep learning schemes. Multimed Tools Appl. 2024;83:83157-77. [Crossref]
- Keshamoni K, Rao LK, Rao DS. A multi-modal CNN-based approach for COVID-19 diagnosis using ECG, X-Ray, and CT. International Journal of Advanced Computer Science&Applications (IJACSA). 2024;15(6):1102-12. [Crossref]
- Mousavi M, Hosseini S. A deep convolutional neural network approach using medical image classification. BMC Med Inform Decis Mak. 2024;24:239. [Crossref]
- Rajpoot R, Gour M, Jain S, Semwal VB. Integrated ensemble CNN and explainable AI for COVID-19 diagnosis from CT scan and X-ray images. Sci Rep. 2024;14(1):24985. [Crossref] [PubMed] [PMC]
- Tello-Mijares S, Flores F, Woo F. Classification of severity of lung parenchyma using saliency and discrete cosine transform energy in computed tomography of patients with COVID-19. Int J Telemed Appl. 2025;2025:4420410. [Crossref] [PubMed] [PMC]
.: İşlem Listesi