Unknown

Dataset Information

0

Severity Classification of Conjunctival Hyperaemia by Deep Neural Network Ensembles.


ABSTRACT: Conjunctival hyperaemia is a common clinical ophthalmological finding and can be a symptom of various ocular disorders. Although several severity classification criteria have been proposed, none include objective severity criteria. Neural networks and deep learning have been utilised in ophthalmology, but not for the purpose of classifying the severity of conjunctival hyperaemia objectively. To develop a conjunctival hyperaemia grading software, we used 3700 images as the training data and 923 images as the validation test data. We trained the nine neural network models and validated the performance of these networks. We finally chose the best combination of these networks. The DenseNet201 model was the best individual model. The combination of the DenseNet201, DenseNet121, VGG19, and ResNet50 were the best model. The correlation between the multimodel responses, and the vessel-area occupied was 0.737 (p < 0.01). This system could be as accurate and comprehensive as specialists but would be significantly faster and consistent with objective values.

SUBMITTER: Masumoto H 

PROVIDER: S-EPMC6589312 | biostudies-literature | 2019

REPOSITORIES: biostudies-literature

altmetric image

Publications

Severity Classification of Conjunctival Hyperaemia by Deep Neural Network Ensembles.

Masumoto Hiroki H   Tabuchi Hitoshi H   Yoneda Tsuyoshi T   Nakakura Shunsuke S   Ohsugi Hideharu H   Sumi Tamaki T   Fukushima Atsuki A  

Journal of ophthalmology 20190602


Conjunctival hyperaemia is a common clinical ophthalmological finding and can be a symptom of various ocular disorders. Although several severity classification criteria have been proposed, none include objective severity criteria. Neural networks and deep learning have been utilised in ophthalmology, but not for the purpose of classifying the severity of conjunctival hyperaemia objectively. To develop a conjunctival hyperaemia grading software, we used 3700 images as the training data and 923 i  ...[more]

Similar Datasets

| S-EPMC7104788 | biostudies-literature
| S-EPMC11375051 | biostudies-literature
| S-EPMC8238938 | biostudies-literature
| S-EPMC5537102 | biostudies-other
| S-EPMC7492909 | biostudies-literature
| S-EPMC7214032 | biostudies-literature
| S-EPMC6186044 | biostudies-literature
| S-EPMC10272759 | biostudies-literature
| S-EPMC7961024 | biostudies-literature
| S-EPMC7071616 | biostudies-literature