Göz Katarakt Hastalığının Tanısında Özellik Çıkarma Aşaması Olarak Transfer Öğrenme Tekniğinin Kullanımı

Yıl 2022, Cilt: 1 Sayı: 1, 17 - 33, 18.08.2022

Halit Bakır Şahin Yılmaz

Öz

Dünya Sağlık Örgütü'nün (WHO) verilerine göre şu anda dünya çapında en az 2,2 milyar insan görme bozukluğuna sahiptir ve bunlardan en az 1 milyarı önlenebilir görme bozukluğuna sahiptir. Göz hastalıkları dünya çapında özellikle de gelişmekte olan ve az gelişmiş ve ülkelerde ciddi bir sorun haline gelmiştir. Gündelik yaşamın sürdürülmesinde en önemli organlardan biri olan göz, hayati bir öneme sahiptir. Erken oküler hastalık tespiti, göz hastalıklarının sebep olduğu körlüğü ortadan kaldırmada etkili ve ekonomik bir yöntemdir. Bu çalışmada, retinal fundus görüntülerinden katarakt hastalığı tespit etmek için bir derin öğrenme modeli önerilmiştir. Önerilen model iki fazdan oluşmaktadır. Birinci fazda, VGG-16, ResNet, InceptionV3 ve MobileNet gibi bazı meşhur evrişimli sinir ağı mimarilerin özellik ayıklama faz olarak kullanılması önerilmiştir. İkinci fazda ise, klasik sinir ağ katmanları kullanılarak sınıflandırma işlemi gerçekleştirilmiştir. Önerilen model, 8 sınıftan 6392 görüntü içeren retina görüntü veri setinden iki sınıflı bir veri seti oluşturularak test edilmiştir. Önerilen modelde, özellik ayıklamak için ResNet kullanıldığında %95.51 gibi yüksek bir oranda tespit doğruluğuna ulaşılmıştır. Önerilen yöntemin, katarakt hastalığının teşhisinde büyük ölçüde etkili ve başarılı olduğu görülmüş olup, bu yöntem kullanılarak diğer tüm göz hastalıklarını teşhis etmede kullanılabileceği düşünülmektedir.

Anahtar Kelimeler

Katarakt, Oftalmoloji, Fundus Görüntüleri, Transfer Öğrenme, Evrişimli Sinir Ağ

Using Transfer Learning Technique as a Feature Extraction Phase for Diagnosis of Cataract Disease in the Eye

Öz

According to the data of the World Health Organization (WHO), currently at least 2.2 billion people worldwide have visual impairment, and at least 1 billion of them have preventable visual impairment. Eye diseases have become a serious problem, especially in the developing and underdeveloped countries around the world. The eye is one of the most important organs in the maintenance of daily life. So, early detection of ocular disease is an effective and economical method of eliminating blindness caused by eye diseases. In this work, a deep learning model has been proposed for detecting the cataract disease from retinal fundus images. The proposed model consists of two phases. In the first phase, it is proposed to use some famous convolutional neural network architectures such as VGG-16, ResNet, Inception v3 and MobileNet as a feature extraction phase. In the second phase, some classical neural network layers have been adopted and trained using the features extracted in the first phase for conducting the classification process. The proposed model has been trained and tested using a dataset contains two classes selected from a retinal image dataset containing 6392 images related to 8 classes. The proposed model gave high detection accuracy, where the best results reached 95.51%, which has been obtained when the ResNet well-known deep learning model has been used as feature extraction phase in the proposed model. The proposed method has shown that it is largely effective and successful in the diagnosis of cataract disease, and it can be generalized to be used for diagnosing all eye diseases.

Anahtar Kelimeler

Cataract, ophthalmology, fundus images, transfer learning, convolutional neural network

Kaynakça

• Referans1:Abràmoff, Michael David et al. 2016. “Improved Automated Detection of Diabetic Retinopathy on a Publicly Available Dataset through Integration of Deep Learning.” Investigative ophthalmology & visual science 57(13): 5200–5206.
• Referans2:Burlina, Philippe et al. 2016. “Detection of Age-Related Macular Degeneration via Deep Learning.” In 2016 IEEE 13th International Symposium on Biomedical Imaging (ISBI), IEEE, 184–88.
• Referans3:Grassmann, Felix et al. 2018. “A Deep Learning Algorithm for Prediction of Age-Related Eye Disease Study Severity Scale for Age-Related Macular Degeneration from Color Fundus Photography.” Ophthalmology 125(9): 1410–20.
• Referans4:Gulshan, Varun et al. 2016. “Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs.” Jama 316(22): 2402–10.
• Referans5:He, Kaiming, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. “Deep Residual Learning for Image Recognition.” In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, , 770–78.
• Referans6:Howard, Andrew G et al. 2017. “Mobilenets: Efficient Convolutional Neural Networks for Mobile Vision Applications.” arXiv preprint arXiv:1704.04861.
• Referans7:Li, Feng et al. 2019. “Fully Automated Detection of Retinal Disorders by Image-Based Deep Learning.” Graefe’s Archive for Clinical and Experimental Ophthalmology 257(3): 495–505.
• Referans8:Peng, Yifan et al. 2019. “DeepSeeNet: A Deep Learning Model for Automated Classification of Patient-Based Age-Related Macular Degeneration Severity from Color Fundus Photographs.” Ophthalmology 126(4): 565–75.
• Referans9:Poly, Tahmina Nasrin et al. 2019. “Artificial Intelligence in Diabetic Retinopathy: Insights from a Meta-Analysis of Deep Learning.” In MEDINFO 2019: Health and Wellbeing e-Networks for All, IOS Press, 1556–57.
• Referans10:Poplin, Ryan et al. 2018. “Prediction of Cardiovascular Risk Factors from Retinal Fundus Photographs via Deep Learning.” Nature Biomedical Engineering 2(3): 158–64.
• Referans11:Porumb, Mihaela, Saverio Stranges, Antonio Pescapè, and Leandro Pecchia. 2020. “Precision Medicine and Artificial Intelligence: A Pilot Study on Deep Learning for Hypoglycemic Events Detection Based on ECG.” Scientific reports 10(1): 1–16.
• Referans12:Shin, Seung Yeon, Soochahn Lee, Il Dong Yun, and Kyoung Mu Lee. 2019. “Deep Vessel Segmentation by Learning Graphical Connectivity.” Medical image analysis 58: 101556.
• Referans13:Simonyan, Karen, and Andrew Zisserman. 2014. “Very Deep Convolutional Networks for Large-Scale Image Recognition.” arXiv preprint arXiv:1409.1556.
• Referans14:Szegedy, Christian et al. 2015. “Going Deeper with Convolutions.” In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1–9.
• Referans15:Ting, Daniel Shu Wei et al. 2017. “Development and Validation of a Deep Learning System for Diabetic Retinopathy and Related Eye Diseases Using Retinal Images from Multiethnic Populations with Diabetes.” Jama 318(22): 2211–23.
• Referans16:Tufail, Adnan et al. 2017. “Automated Diabetic Retinopathy Image Assessment Software: Diagnostic Accuracy and Cost-Effectiveness Compared with Human Graders.” Ophthalmology 124(3): 343–51.
• Referans17:WHO TEAM. 2019. World Report on Vision.
• Referans18:Yang, Ji-Jiang et al. 2016. “Exploiting Ensemble Learning for Automatic Cataract Detection and Grading.” Computer methods and programs in biomedicine 124: 45–57.
• Referans19:Yang, Meimei et al. 2013. “Classification of Retinal Image for Automatic Cataract Detection.” In 2013 IEEE 15th International Conference on E-Health Networking, Applications and Services (Healthcom 2013), IEEE, 674–79.