Bayesian Optimized Machine Learning Model for Automated Eye Disease Classification from Fundus Images

被引:2
作者
Zannah, Tasnim Bill [1 ]
Abdulla-Hil-Kafi, Md. [1 ]
Sheakh, Md. Alif [1 ]
Hasan, Md. Zahid [1 ]
Shuva, Taslima Ferdaus [1 ]
Bhuiyan, Touhid [1 ]
Rahman, Md. Tanvir [2 ,3 ]
Khan, Risala Tasin [4 ]
Kaiser, M. Shamim [4 ]
Whaiduzzaman, Md [5 ,6 ]
机构
[1] Daffodil Int Univ, Dept Comp Sci & Engn, Hlth Informat Res Lab HIRL, Dhaka 1341, Bangladesh
[2] Univ Queensland, Sch Hlth & Rehabil Sci, Brisbane, Qld 4072, Australia
[3] Mawlana Bhashani Sci & Technol Univ, Dept Informat & Commun Technol, Tangail 1902, Bangladesh
[4] Jahangirnagar Univ, Inst Informat Technol, Dhaka 1342, Bangladesh
[5] Torrens Univ, Ctr Artificial Intelligence Res & Optimisat AIRO, Adelaide, SA 5000, Australia
[6] Queensland Univ Technol, Sch Informat Syst, Brisbane, Qld 4000, Australia
关键词
eye disease; machine learning; principal component analysis; Bayesian optimization;
D O I
10.3390/computation12090190
中图分类号
O1 [数学];
学科分类号
0701 ; 070101 ;
摘要
Eye diseases are defined as disorders or diseases that damage the tissue and related parts of the eyes. They appear in various types and can be either minor, meaning that they do not last long, or permanent blindness. Cataracts, glaucoma, and diabetic retinopathy are all eye illnesses that can cause vision loss if not discovered and treated early on. Automated classification of these diseases from fundus images can empower quicker diagnoses and interventions. Our research aims to create a robust model, BayeSVM500, for eye disease classification to enhance medical technology and improve patient outcomes. In this study, we develop models to classify images accurately. We start by preprocessing fundus images using contrast enhancement, normalization, and resizing. We then leverage several state-of-the-art deep convolutional neural network pre-trained models, including VGG16, VGG19, ResNet50, EfficientNet, and DenseNet, to extract deep features. To reduce feature dimensionality, we employ techniques such as principal component analysis, feature agglomeration, correlation analysis, variance thresholding, and feature importance rankings. Using these refined features, we train various traditional machine learning models as well as ensemble methods. Our best model, named BayeSVM500, is a Support Vector Machine classifier trained on EfficientNet features reduced to 500 dimensions via PCA, achieving 93.65 +/- 1.05% accuracy. Bayesian hyperparameter optimization further improved performance to 95.33 +/- 0.60%. Through comprehensive feature engineering and model optimization, we demonstrate highly accurate eye disease classification from fundus images, comparable to or superior to previous benchmarks.
引用
收藏
页数:18
相关论文
共 46 条
[1]   EfficientNet family U-Net models for deep learning semantic segmentation of kidney tumors on CT images [J].
Abdelrahman, Abubaker ;
Viriri, Serestina .
FRONTIERS IN COMPUTER SCIENCE, 2023, 5
[2]  
Ahmed MR, 2021, Academic Platform Journal of Engineering and Smart Systems, V9, P47, DOI [10.21541/apjes.741194, 10.21541/apjes.741194, DOI 10.21541/APJES.741194]
[3]  
Ambarwati Yulia Siti, 2020, 2020 3rd International Seminar on Research of Information Technology and Intelligent Systems (ISRITI), P694, DOI 10.1109/ISRITI51436.2020.9315486
[4]  
[Anonymous], 2022, Kaggle
[5]  
[Anonymous], 2022, Glaucoma Detection.
[6]  
[Anonymous], 2019, KAGGLE
[7]  
Babaqi T, 2023, Arxiv, DOI arXiv:2307.10501
[8]   Promising Artificial Intelligence-Machine Learning-Deep Learning Algorithms in Ophthalmology [J].
Balyen, Lokman ;
Peto, Tunde .
ASIA-PACIFIC JOURNAL OF OPHTHALMOLOGY, 2019, 8 (03) :264-272
[9]   Classification of Eye Diseases in Fundus Images [J].
Bernabe, Omar ;
Acevedo, Elena ;
Acevedo, Antonio ;
Carreno, Ricardo ;
Gomez, Sandra .
IEEE ACCESS, 2021, 9 :101267-101276
[10]   Automatic detection of 39 fundus diseases and conditions in retinal photographs using deep neural networks [J].
Cen, Ling-Ping ;
Ji, Jie ;
Lin, Jian-Wei ;
Ju, Si-Tong ;
Lin, Hong-Jie ;
Li, Tai-Ping ;
Wang, Yun ;
Yang, Jian-Feng ;
Liu, Yu-Fen ;
Tan, Shaoying ;
Tan, Li ;
Li, Dongjie ;
Wang, Yifan ;
Zheng, Dezhi ;
Xiong, Yongqun ;
Wu, Hanfu ;
Jiang, Jingjing ;
Wu, Zhenggen ;
Huang, Dingguo ;
Shi, Tingkun ;
Chen, Binyao ;
Yang, Jianling ;
Zhang, Xiaoling ;
Luo, Li ;
Huang, Chukai ;
Zhang, Guihua ;
Huang, Yuqiang ;
Ng, Tsz Kin ;
Chen, Haoyu ;
Chen, Weiqi ;
Pang, Chi Pui ;
Zhang, Mingzhi .
NATURE COMMUNICATIONS, 2021, 12 (01)