【医学+深度论文:F11】2018 A deep learning model for the detection of both advanced and early glaucoma using
11
2018 PLOS
A deep learning model for the detection of both advanced and early glaucoma using fundus photography
Method : 分类 正常眼 、青光眼(早期晚期)
Dataset : 1542 (786 N,467 AG,289 EG)
AG : advanced glaucoma
EG : early glaucoma
非散瞳自动眼底照相机(AFC-330,Nidek,Japan)
Architecture : CNN(3con+3maxpooling+2FC)
Results :accuracy 87.9% AUROC 0.94
Methods
Dataset 1542
- training 754
- validation 324
- test 464
三种method对比实验
- logistics classification
1×(240×240×3)的一维阵列 → softmax → 和one-hot encoded labels 比较
→ 交叉熵计算损失
-
transfer GoogleNet Inception v3
Data augmentation
10 images per photography -
transfer
pretraining -
GoogleNet Inception v3
修改了Inception v3模型的最后一个分类层以满足我们的分类需求,然后使用我们的数据进行微调
our CNN
-
Data augmentation
10 images per photography
从单个图像生成10个概率。通过平均该概率,我们可以基于每个图像获得图像代表青光眼或正常患者的单一概率
Results
-
logistics classification : accuracy 77.2%
-
transfer 的GoogleNet Inception v3 :accuracy 84.5% AUROC 0.93
-
out model :accuracy 87.9% AUROC 0.94
Discussion
-
传统的机器学习方法如逻辑分类足以对晚期青光眼进行分类。但是对晚期和早期青光眼检测还需要深度学习
-
深度学习可能会过拟合,所以用了数据增强和dropout
-
解释为什么网络结构简单,说的是因为是为了二分类设计的。后期可能继续优化模型
-
做了 深度学习和传统机器学习比较以及自己的模型和 GoogleNet Inception v3 的比较,证明了深度学习比机器学习在青光眼分类的精度上更好(能很好地分类早期青光眼),自己的模型比经过transfer 的GoogleNet Inception v3更好。还证明了数据增强和dropout是由必要的。
但是我认为对于证明机器学习没有深度学习好这块实验仅仅用了一个简单的逻辑回归,不足以说明。 - 搞不明白为什么 GoogleNet Inception v3 为什么会不如他的模型
当验证数据的每个时期的平均损失开始增加时,认为训练阶段结束。这个是不是不妥。训练方式有没有问题。
- 论文笔记 A Large Contextual Dataset for Classification,Detection and Counting of Cars with Deep Learning
- 【医学+深度论文:S01】Retinal Fundus Image for Glaucoma Detection: A Review and Study
- [深度学习论文笔记][Weight Initialization] Exact solutions to the nonlinear dynamics of learning in deep lin
- Automatic Feature Learning for Glaucoma Detection Based on Deep Learning论文理解
- 深度学习笔记——“Mastering the game of Go with deep neural networks and tree search”论文学习
- 论文笔记《A Survey of Model Compression and Acceleration for Deep Neural Networks》
- 论文学习-深度学习目标检测2014至201901综述-Deep Learning for Generic Object Detection A Survey
- 论文复现报告:Deep Region and Multi-label Learning for Facial Action Unit Detection
- Deep Region and Multi-label Learning for Facial Action Unit Detection简要论文笔记
- 【医学+深度论文:F14】2018 Joint Optic Disc and Cup Segmentation Based on Multi Label Deep Network
- 实时SLAM的未来及与深度学习的比较The Future of Real-Time SLAM and “Deep Learning vs SLAM”
- 【医学+深度论文:F12】2017Glaucoma-Deep: Detection of Glaucoma Eye Disease on Retinal Fundus Images using Dee
- Joint Deep Learning For Pedestrian Detection(论文笔记-深度学习:行人检测)
- 论文阅读:End-to-End Learning of Deformable Mixture of Parts and Deep Convolutional Neural Networks for H
- The Wide and Deep Learning Model(译文+Tensorlfow源码解析) 原创 2017年11月03日 22:14:47 标签: 深度学习 / 谷歌 / tensorf
- 《The Frontiers of Memory and Attention in Deep Learning》 图文结合详解深度学习Memory & Attention
- Deep Learning for Design and Retr of Nano-photonic Structures 论文学习
- [论文笔记] Leveraging the crowd as a source of innovation Does crowdsourcing represent a new model for product and service innovation? (SIGMIS-CPR, 2012)
- 【医学+深度论文:F15】2018 Disc-Aware Ensemble Network for Glaucoma Screening From Fundus Image
- 论文笔记:Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks