Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?

卷积神经网络 计算机科学 人工智能 深度学习 刮擦 微调 背景(考古学) 分割 模式识别(心理学) 图像分割 学习迁移 图像(数学) 机器学习 医学影像学 上下文图像分类 古生物学 物理 量子力学 生物 操作系统
作者
Nima Tajbakhsh,J. Shin,Suryakanth Gurudu,R. Todd Hurst,Christopher B. Kendall,Michael B. Gotway,Jianming Liang
出处
期刊:IEEE Transactions on Medical Imaging [Institute of Electrical and Electronics Engineers]
卷期号:35 (5): 1299-1312 被引量:3073
标识
DOI:10.1109/tmi.2016.2535302
摘要

Training a deep convolutional neural network (CNN) from scratch is difficult because it requires a large amount of labeled training data and a great deal of expertise to ensure proper convergence. A promising alternative is to fine-tune a CNN that has been pre-trained using, for instance, a large set of labeled natural images. However, the substantial differences between natural and medical images may advise against such knowledge transfer. In this paper, we seek to answer the following central question in the context of medical image analysis: Can the use of pre-trained deep CNNs with sufficient fine-tuning eliminate the need for training a deep CNN from scratch? To address this question, we considered four distinct medical imaging applications in three specialties (radiology, cardiology, and gastroenterology) involving classification, detection, and segmentation from three different imaging modalities, and investigated how the performance of deep CNNs trained from scratch compared with the pre-trained CNNs fine-tuned in a layer-wise manner. Our experiments consistently demonstrated that 1) the use of a pre-trained CNN with adequate fine-tuning outperformed or, in the worst case, performed as well as a CNN trained from scratch; 2) fine-tuned CNNs were more robust to the size of training sets than CNNs trained from scratch; 3) neither shallow tuning nor deep tuning was the optimal choice for a particular application; and 4) our layer-wise fine-tuning scheme could offer a practical way to reach the best performance for the application at hand based on the amount of available data.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
畅快盼望发布了新的文献求助10
刚刚
博雯发布了新的文献求助30
1秒前
ann完成签到,获得积分10
2秒前
坚强书文发布了新的文献求助10
2秒前
123456完成签到 ,获得积分10
3秒前
3秒前
Hey完成签到 ,获得积分10
3秒前
科目三应助cata采纳,获得10
3秒前
bkagyin应助雷豪采纳,获得10
4秒前
5秒前
5秒前
打打应助gouqi采纳,获得10
5秒前
5秒前
6秒前
7秒前
8秒前
yzy发布了新的文献求助10
8秒前
9秒前
梁队长发布了新的文献求助10
10秒前
10秒前
12秒前
所所应助傻子与白痴采纳,获得10
12秒前
14秒前
14秒前
无限达完成签到,获得积分10
14秒前
16秒前
科研通AI6.4应助畅快盼望采纳,获得10
17秒前
坚强书文完成签到,获得积分10
18秒前
ke完成签到,获得积分10
18秒前
19秒前
20秒前
21秒前
22秒前
打打应助sai采纳,获得10
22秒前
木木发布了新的文献求助10
23秒前
iris完成签到,获得积分10
24秒前
25秒前
南极以南发布了新的文献求助10
25秒前
务实寻真发布了新的文献求助10
26秒前
woshizy完成签到,获得积分10
27秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
2026年中国辛酸癸酸聚乙二醇甘油酯行业市场现状调查及投资机会研判报告 1000
2026年中国辛酸癸酸聚乙二醇甘油酯行业市场规模及竞争格局分析报告 1000
Resiliency Scale for Adolescents--Chinese Version 800
Fundamentals of Pharmaceutical and Biologics Regulations: A Global Perspective, Second Edition 700
作者名:Kristopher P. Plain,悉尼大学的,目前只能查到其四篇论文,想找到其博士论文 550
Matrix Methods in Data Mining and Pattern Recognition Second Edition 510
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 工程类 有机化学 化学工程 生物化学 计算机科学 内科学 物理 复合材料 催化作用 细胞生物学 无机化学 光电子学 物理化学 电极 基因
热门帖子
关注 科研通微信公众号,转发送积分 7328628
求助须知:如何正确求助?哪些是违规求助? 8943260
关于积分的说明 18969254
捐赠科研通 6984352
什么是DOI,文献DOI怎么找? 3216357
关于科研通互助平台的介绍 2383041
邀请新用户注册赠送积分活动 2195805