DNN Classifier of Wide-angle Retinal Images in Computer-Aided Screening for ROP

This repository contains the following files:<br>1. data set preview.png<br>A preview of the data set. This is a snapshot of the resized (80x60) version of the training &amp; test data set.The original image data set are not provided, due to patient privacy protection and hospital IRB regulations.<br>2. Training Set Thumbnails.zip<br>The training dataset includes:Class I (with disease): 8089Class II (without disease): 9711<br><br>3. Test Set Thumbnails.zip<br>The test training dataset includes:Class I (with disease): 155Class II (without disease): 1587<br>4. DNNs.zip<br>This zip file contains 3 subfolders: AlexNet, GoogLeNet, VGG16.Each subfolder further includes: Model definition graphTwo trained models in the Caffe format. One is trained "from scratch" and the other "by transfer learning". P-R curveROC curveSoftmax output probabilities on the test data set<br><br>5. HumanExpertsPerformance.zip<br>The five pediatric ophthalmologists' diagnoses on the test data setDNN performance measures<br>DNN vs Human P-R curveDNN vs Human ROC curve<br><br>6. scripts.zip<br>Scripts used in this study for batch inference and data analysis. <br><br>

本仓库包含以下文件：<br>1. data set preview.png：数据集预览图。该图为训练与测试数据集经调整尺寸至80×60后的快照。由于患者隐私保护及医院伦理审查委员会（IRB）相关规定，未提供原始图像数据集。<br>2. Training Set Thumbnails.zip：训练集缩略压缩包。训练数据集包含：类别I（Class I，伴发病变）：8089例；类别II（Class II，无病变）：9711例。<br>3. Test Set Thumbnails.zip：测试集缩略压缩包。测试数据集包含：类别I（Class I，伴发病变）：155例；类别II（Class II，无病变）：1587例。<br>4. DNNs.zip：深度神经网络（DNNs，Deep Neural Networks）压缩包。该压缩包包含3个子文件夹：AlexNet、GoogLeNet、VGG16。每个子文件夹均包含以下内容：模型定义图、两种Caffe格式的训练模型（一种为从头开始训练，另一种采用迁移学习训练）、精确召回（P-R，Precision-Recall）曲线、受试者工作特征（ROC，Receiver Operating Characteristic）曲线、测试集上的Softmax输出概率。<br>5. HumanExpertsPerformance.zip：人类专家性能压缩包。该压缩包包含5名儿科眼科医师对测试数据集的诊断结果、深度神经网络性能评估指标、深度神经网络与人类专家的P-R曲线、深度神经网络与人类专家的ROC曲线。<br>6. scripts.zip：脚本压缩包。本研究中用于批量推理与数据分析的脚本文件。