Rapid classification of glaucomatous fundus images
Rapid classification of glaucomatous fundus images
复制标题
DOI：
10.1364/josaa.415395
复制
发表时间：
2021-06-01
影响因子：
1.9
通讯作者：
Lakshminarayanan, Vasudevan
中科院分区：
文献类型：
--
作者：
Singh, Hardit;Saini, Simarjeet S.;Lakshminarayanan, Vasudevan
本文提出了一种新的卷积神经网络训练方法，并利用卷积神经网络对眼底图像进行青光眼分类。该方法集成了强化学习和监督学习，并将其用于迁移学习。该训练方法通过“随机运动”和“随机检测”两种不同的攀岩者类型来使用爬山技术，并通过随机梯度下降动量模型与监督学习模型相结合。使用具有青光眼和正常眼底图像的Drishti-GS和RIM-ONE-r2数据集对模型进行训练和测试。在GoogLeNet、DenseNet-201、NASNet、VGG-19和Inception-Resnet v2这五个CNN架构上通过迁移学习测试了预测性能。使用五重分类来评估性能，在保持高精度的同时实现了高灵敏度。在测试的模型中，DenseNet-201架构在灵敏度和曲线下面积方面表现最好。这种训练方法允许在小数据集上进行迁移学习，并且可以应用于远程眼科应用，包括使用本地数据集进行训练。(C) 2021美国光学学会
We propose a new method for training convolutional neural networks (CNNs) and use it to classify glaucoma from fundus images. This method integrates reinforcement learning along with supervised learning and uses it for transfer learning. The training method uses hill climbing techniques via two different climber types, namely, "random movement" and "random detection," integrated with a supervised learning model through a stochastic gradient descent with momentum model. The model was trained and tested using the Drishti-GS and RIM-ONE-r2 datasets having glaucomatous and normal fund us images. The performance for prediction was tested by transfer learning on five CNN architectures, namely, GoogLeNet, DenseNet-201, NASNet, VGG-19, and Inception-Resnet v2. A five-fold classification was used for evaluating the performance, and high sensitivities while maintaining high accuracies were achieved. Of the models tested, the DenseNet-201 architecture performed the best in terms of sensitivity and area under the curve. This method of training allows transfer learning on small datasets and can be applied for tele-ophthalmology applications including training with local datasets. (C) 2021 Optical Society of America