Effect of color information on the diagnostic performance of glaucoma in deep learning using few fundus images
- PDF / 1,077,227 Bytes
- 10 Pages / 547.087 x 737.008 pts Page_size
- 48 Downloads / 168 Views
(0123456789().,-volV) ( 01234567 89().,-volV)
ORIGINAL PAPER
Effect of color information on the diagnostic performance of glaucoma in deep learning using few fundus images Masakazu Hirota . Atsushi Mizota . Tatsuya Mimura . Takao Hayashi . Junichi Kotoku . Tomohiro Sawa . Kenji Inoue
Received: 14 December 2019 / Accepted: 20 June 2020 Ó Springer Nature B.V. 2020
Abstract Purpose The purpose of this study was to evaluate the accuracy of the convolutional neural network (CNN) model in glaucoma identification with three primary colors (red, green, blue; RGB) and split color channels using fundus photographs with a small sample size. Methods The dataset was prepared using color fundus photographs captured with a fundus camera (VX-10i, Kowa Co., Ltd., Tokyo, Japan). The training dataset consisted of 200 images, and the validation
M. Hirota T. Hayashi Department of Orthoptics, Faculty of Medical Technology, Teikyo University, Itabashi, Tokyo, Japan M. Hirota (&) A. Mizota T. Mimura T. Hayashi Department of Ophthalmology, School of Medicine, Teikyo University, 2-11-1 Kaga, Itabashi-ku, Tokyo 173-8605, Japan e-mail: [email protected] J. Kotoku Division of Clinical Radiology, Graduate School of Medical Care and Technology, Teikyo University, Itabashi, Tokyo, Japan
dataset contained 60 images. In the preprocessing stage, the color channels for the fundus images were separated into red (red model), green (green model), and blue (blue model) using OpenCV on Windows. All images were resized to squares with a size of 512 9 512 pixels for preprocessing before input into the model, and the model was fine-tuned with VGG16. Results The diagnostic performance was significantly higher in the green model [area under the curve (AUC) 0.946; 95% confidence interval (CI) 0.851–0.982] than in the RGB model (AUC 0.800; 95% CI 0.658–0.893; P = 0.006), red model (AUC 0.746; 95% CI 0.601–0.851; P = 0.002), and blue model (AUC 0.558; 95% CI 0.405–0.700; P \ 0.001). Conclusion The present study showed that the green digital filter is useful for structuring CNN models for automatic discrimination of glaucoma using fundus photographs with a small sample size. The present findings suggest that preprocessing, when creating the CNN model, is an important step for the identification of a large number of retinal diseases using color fundus photographs. Keywords Glaucoma Fundus photography Deep learning Convolutional neural network
T. Sawa Medical Information Systems Research Center, Teikyo University, Itabashi, Tokyo, Japan K. Inoue Inoue Eye Hospital, Chiyoda, Tokyo, Japan
123
Int Ophthalmol
Introduction Glaucoma is an irreversible disease and is the second most common cause of blindness worldwide [1]. A recent meta-analysis based on 50 glaucoma prevalence-related studies reported that the prevalence of glaucoma was 3.5% among individuals aged 40–80 years, 3.0% for primary open-angle glaucoma, and 0.5% for primary angle closure glaucoma, corresponding to an estimated 64.3 million individuals with the disease world
Data Loading...
