fundus_binary_classification 眼底图二分类——densenet 40层实现 图像对比 正样本/负样本 Dependencies keras 其中,训练集正负样本各约3600(data/train),验证集各约900(data/validation)。输入图像大小32×32×3 训练: python densenet_40.py result after epoch=10, accuracy = 77% #if increase the epoch,the accuracy may higher. 训练准确率、损失 验证准确率、损失