Hello World Deep Learning in Medical Imaging

There is recent popularity in applying machine learning to medical imaging, notably deep learning, which has achieved state-of-the-art performance in image analysis and processing. The rapid adoption of deep learning may be attributed to the availability of machine learning frameworks and libraries to simplify their use. In this tutorial, we provide a high-level overview of how to build a deep neural network for medical image classification, and provide code that can help those new to the field begin their informatics projects.     

深度学习在脑胶质瘤影像学的研究进展

深度学习是基于多层神经网络计算模型发现数据内复杂特征的一种深度网络,较多应用于医学图像的分割与分类中,在各类脑胶质瘤的研究中也有许多成果。本文就深度学习在脑胶质瘤的准确分割定位、组织遗传学特征预测及预后评估等方面展开综述,总结深度学习在脑胶质瘤影像图像分割与分类的研究进展,从而为胶质瘤患者的精准诊断、个体化治疗提供新思路。     

医学图像语义概念识别方法研究

医学图像语义概念识别是医学图像知识表示的重要技术环节。研究医学图像语义概念识别方法,有助于机器理解和学习医学图像中的潜在医学知识,在影像辅助诊断和智能读片等应用中发挥重要作用。将医学图像的高频概念识别问题转化为多标签分类任务,利用基于卷积神经网络的深度迁移学习方法,识别有限数量的高频医学概念;同时利用基于图像检索的主题建模方法,从给定医学图像的相似图像中提取语义相关概念。国际跨语言图像检索论坛ImageCLEF于2018年5月组织ImageCLEFcaption 2018评测,其子任务“概念检测”的目标是给定222 314张训练图片和9 938张测试图片,识别111 156个语义概念。上述两种方法的实验结果已被提交。实验结果表明,利用基于卷积神经网络的深度迁移学习方法识别医学图像高频概念,F1值为0.092 8,在提交团队中排名第二;基于图像检索的主题模型可召回部分低频相关概念,F1值为0.090 7,然而其性能依赖于图像检索结果的质量。基于卷积神经网络的深度迁移学习方法识别医学图像高频概念的鲁棒性优于基于图像检索方法的鲁棒性,但在大规模开放语义概念的识别技术研究上仍需进一步完善。     

基于深度学习的多模态医学图像融合方法研究进展

医学图像融合方法可以将有用的信息整合到一张图上，提高单张图像的信息量。对多模态医学图像进行融合时，如何对图像进行有效的变换，提取到不同图像中独有的特征，并施以适当的融合规则是医学图像融合领域研究的重点。近年随着深度学习的快速发展，深度学习被广泛应用于医学图像领域，代替传统方法中的一些人工操作，并在图像表示、图像特征提取以及融合规则的选择方面显示出独特优势。本文针对基于深度学习的医学图像融合进展予以探讨，介绍了卷积神经网络、卷积稀疏表示、深度自编码和深度信念网络这些常用于医学图像融合的框架，对一些应用于融合过程不同步骤的深度学习方法进行分析和总结，最后，分析了当前基于深度学习的融合方法的不足并展望了未来的研究方向。     

Medical image translation using an edge-guided generative adversarial network with global-to-local feature fusion

In this paper, we propose a framework based deep learning for medical image translation using paired and unpaired training data. Initially, a deep neural network with an encoder-decoder structure is proposed for image-to-image translation using paired training data. A multi-scale context aggregation approach is then used to extract various features from different levels of encoding, which are used during the corresponding network decoding stage. At this point, we further propose an edge-guided generative adversarial network for image-to-image translation based on unpaired training data. An edge constraint loss function is used to improve network performance in tissue boundaries. To analyze framework performance, we conducted five different medical image translation tasks. The assessment demonstrates that the proposed deep learning framework brings significant improvement beyond state-of-the-arts.     

基于深度学习的舌象自监督聚类

目的：解决人工智能舌诊领域数据标注成本较高且带有较强主观性的问题。方法：基于深度学习中的对比学习技术,对舌象进行自监督聚类。该方法首先利用卷积神经网络将不同数据增强模式下的舌象映射到潜在空间,并在学习同类实例之间共同特征的同时尽可能区分非同类实例;随后利用高斯混合模型对网络提取的特征向量进行聚类。结果：在无需引入先验知识的情况下,利用300张舌象仪采集的无标签图像取得了52.54%的聚类纯度。结论：该方法一定程度上将医疗工作者从费事费力的数据标注工作中解放出来。除应用于自动化舌象分类外,该方法还可进一步针对不同病症的特殊舌象症候群进行聚类分析,其提取的舌象特征也可为舌体分割、舌色分类、苔质分区等下游任务提供预训练的参考。     

基于深度残差网络的皮肤镜图像黑色素瘤的识别

恶性黑色素瘤是最常见和最致命的皮肤癌之一。临床上,皮肤镜检查是恶性黑色素瘤早期诊断的常规手段。但是人工检查费力、费时,并且高度依赖于皮肤科医生的临床经验。因此,研究出自动识别皮肤镜图像中的黑色素瘤算法显得尤为重要。提出一种皮肤镜图像自动评估的新框架,利用深度学习方法,使其在有限的训练数据下产生更具区分性的特征。具体来说,首先在大规模自然图像数据集上预训练一个深度为152层的残差神经网络(Res-152),用来提取皮肤病变图像的深度卷积层特征,并对其使用均值池化得到特征向量,然后利用支持向量机(SVM)对提取的黑色素瘤特征进行分类。在公开的皮肤病变图像ISBI 2016挑战数据集中,用所提出的方法对248幅黑色素瘤图像和1 031幅非黑色素瘤图像进行评估,达到86.28%的准确率及84.18%的AUC值。同时,为论证神经网络深度对分类结果的影响,比较不同深度的模型框架。与现有使用传统手工特征的研究(如基于密集采样SIFT描述符的词袋模型)相比,或仅从深层神经网络的全连接层提取特征进行分类的方法相比,新方法能够产生区分性能更强的特征表达,可以在有限的训练数据下解决黑色素瘤的类内差异大、黑色素瘤与非黑素瘤之间的类间差异小的问题。     

卷积神经网络在医学图像分割中的研究进展

卷积神经网络（CNN）是目前计算机视觉和模式识别中效果最为突出的算法。CNN拥有强大的空间识别能力,可以从图像中提取高阶的空间特征,同时通过共用卷积核的方式大幅减少参数量,从而在提升网络性能的同时保持总参数量在一个合理的、可运算的范畴。部分采用无监督学习的CNN算法可以在没有先验知识的条件下实现一定程度的图像语义分割,大幅减少人工读图的负担。本研究就CNN在医学图像分割中的研究进展和使用CNN时的具体技巧及其效果进行综述。以使用CNN为核心的深度学习工具解决医学图像分割的课题为中心,展示了CNN在有监督学习、半监督学习及无监督学习中的巨大潜力,分析比较了现有方案的优点与不足,探讨了未来CNN在医学图像领域的前进方向。     

基于深度学习的2D/3D医学图像配准研究

2D/3D配准在临床诊断和手术导航规划中有着广泛的应用,可解决医学图像领域中不同维度图像存在信息缺失的问题,能辅助医生在术中精准定位患者的病灶。常规的2D/3D配准方法主要依赖于图像的灰度进行配准,但非常耗时,不利于临床实时性的需求,并且配准过程中容易陷入局部最优值。提出用深度学习的方法来解决2D/3D医学图像配准问题。采用一个基于深度学习的卷积神经网络,通过网络对数字影像重建技术(DRR)进行训练并自动学习图像特征,预测X光图像所对应的参数,从而实现配准。以人体骨盆的模型骨为实验对象,根据骨盆的CT数据生成36000张DRR图像作为训练集,同时通过C臂采集模型骨的50张X光图像作为验证。结果显示,深度学习算法在相关系数、归一化互信息、欧式距离3个精度评价指标上的测试值分别为0.82±0.07、0.32±0.03、61.56±10.91,而常规2D/3D算法对应的测试值分别为0.79±0.07、0.29±0.03、37.92±7.24,说明深度学习算法的配准精度优于常规2D/3D算法的配准精度,且不存在陷入局部最优值的问题。同时,深度学习的配准时间约为0.03s,远低于常规2D/3D配准的时间,可满足临床对于实时配准的需求,未来将进一步开展临床数据的2D/3D配准研究。     

基于Fisher向量和混合描述符的胎盘成熟度分级

胎盘成熟度分级(PMG)对于评估胎儿生长和孕妇健康来说至关重要。目前,PMG主要依赖于临床医生的主观判断,不仅十分耗时,而且由于工作的重复性和冗余性,常会产生误判。传统机器学习中使用的手工特征提取方法,不能很好解决PMG的分级问题,因此提出从B超图像和彩色多普勒能量图像中提取深度混合描述符进行胎盘成熟度自动分级的方法。从深度卷积神经网络中提取卷积特征,并将其与手工特征结合形成混合描述符来提高模型性能。首先,将多个特征层的不同模型进行融合,从图像中获取混合描述符。同时,考虑到深度表达特征,使用迁移学习策略来增强分级性能。然后,用Fisher向量(FV)对提取的描述符进行编码。最后,使用支持向量机(SVM)分类器对胎盘成熟度进行分级。用医生标注好的数据进行测试,在基于19层网络的混合特征模型获得高达94.15%的精确度,比单一使用手工特征模型提升3.01%,比CNN特征模型提升7.35%。实验结果证明,所提方法能够有效应用于胎盘成熟度自动分级。     

医学图像深度学习处理方法的研究进展

由于医学图像数据爆炸式增长,传统依靠医生人工对医学图像进行分析诊断,不仅工作效率低下,工作量大,还容易误诊、漏诊。随着人工智能(artificial intelligence,AI)技术的发展与应用,机器学习(machine learning,ML),尤其是深度学习(deep learning,DL)在医学图像分析领域发挥着越来越重要的作用。本文对DL在医学图像自动分割和分类识别中的研究进展进行综述,为DL在解决医学图像分析诊断方面提供有益参考。     

基于深度学习的医学图像识别研究进展

近年来,随着医学影像技术的快速发展,医学图像分析步入大数据时代,如何从海量的医学图像数据中挖掘出有用信息,对医学图像识别带来巨大的挑战。深度学习是机器学习的一个新领域,传统的机器学习方法不能有效地挖掘到医学图像中蕴含的丰富信息,而深度学习通过模拟人脑建立分层模型,具有强大的自动特征提取、复杂模型构建以及高效的特征表达能力,更重要的是深度学习方法能从像素级的原始数据中逐级提取从底层到高层的特征,这为解决医学图像识别面临的新问题提供了新思路。首先阐述深度学习方法,列举深度学习方法的三种常见的实现模型,并介绍深度学习的训练过程;随后总结了深度学习方法在疾病检测与分类和病变识别两方面的应用情况,以及深度学习应用在医学图像识别中的两个共性问题;最后对深度学习在医学图像识别中存在的问题进行分析及展望.     

深度学习算法在脑电信号解码中的应用

近年来深度学习算法得到飞速发展,在生物医学工程领域的应用也越来越广泛。其中,利用深度学习算法从脑电信号(EEG)中解码生理、心理或病理状态也受到越来越多的关注。综述近年来深度学习算法在EEG解码中的应用,介绍常用算法、典型应用场景、重要进展和现存的问题。首先,论述常用于EEG解码的几类深度学习算法的基本原理,包括卷积神经网络、深度信念网络、自编码器和循环神经网络等。然后,讨论深度学习算法的几个典型EEG解码应用场景,包括脑机接口、情绪与认知识别、疾病辅助诊断。结合应用实例,归纳深度学习算法在EEG解码中的常见问题、解决方案、主要进展和研究趋势。最后,总结深度学习应用于EEG信号解码中仍待解决的一些关键问题,如参数复杂度、训练时间以及泛化能力等。     

基于深度学习网络的医学核磁共振成像超分辨率重构实验

基于深度学习网络的医学核磁共振（MR）图像超分辨重建实验研究,提出并构建一个大规模的高质量用于MR图像超分辨的数据集,涵盖了头颅、膝盖、乳房以及头颈4个部位。通过数据质量筛选和不同低分辨率图像生成方式,在原始图像的高分辨率基础下,以×2、×3、×4的下采样尺度,原始MRI图像形成3种不同尺度下的MR图像数据集,同时给出不同部位超分辨难易程度分析。采用7个在自然图像的超分辨率领域中取得最好效果的深度学习网络,将它们迁移到MR图像中,学习低分辨率MR图像到高低分辨MR图像的映射关系,并对比分析这些深度学习网络在自然图像的超分辨效果。通过实验可以看出,深度学习网络在MR图像超分辨取得了比传统算法更好的效果,部分结果不亚于自然图像;不同部位的超分辨效果差异较大,难以以一个深度学习网络使不同部位均具有更好的超分辨效果。深度学习网络在MR图像超分辨将具有重要的应用价值和理论意义。     

Machine learning for medical diagnosis: history, state of the art and perspective

The paper provides an overview of the development of intelligent data analysis in medicine from a machine learning perspective: a historical view, a state-of-the-art view, and a view on some future trends in this subfield of applied artificial intelligence. The paper is not intended to provide a comprehensive overview but rather describes some subareas and directions which from my personal point of view seem to be important for applying machine learning in medical diagnosis. In the historical overview, I emphasize the naive Bayesian classifier, neural networks and decision trees. I present a comparison of some state-of-the-art systems, representatives from each branch of machine learning, when applied to several medical diagnostic tasks. The future trends are illustrated by two case studies. The first describes a recently developed method for dealing with reliability of decisions of classifiers, which seems to be promising for intelligent data analysis in medicine. The second describes an approach to using machine learning in order to verify some unexplained phenomena from complementary medicine, which is not (yet) approved by the orthodox medical community but could in the future play an important role in overall medical diagnosis and treatment.     

基于浅层与深层特征融合的胃癌前疾病识别

胃癌前疾病识别对降低癌变风险及胃癌发病率具有重要意义。提出一种基于胃镜图像浅层特征与深层特征融合的胃癌前疾病识别方法。首先,根据胃镜图像性质,手工设计75维浅层特征,包含图像的直方图特征、纹理特征以及高阶特征;然后,基于构建的Resnet、GoogLe Net等卷积神经网络,在其输出层前添加一个全连接层作为图像的深层特征,为保证特征权重一致,全连接层的神经元数目设计为75维;最后,串联图像的浅层与深层特征,使用机器学习分类器,识别胃息肉、胃溃疡和胃糜烂等3类胃癌前疾病。对每种疾病收集了380张图像,并以4:1的比例划分为训练集和测试集,然后基于该数据集,分别采用传统机器学习、深度学习、特征融合等3种方法进行模型训练和测试。模型在测试集上的结果显示,所提出的特征融合方法识别准确率高达95.18%,优于传统的机器学习方法(74.12%)和深度学习方法(92.54%)。所提出的方法能够充分利用浅层特征与深层特征,为医生提供临床决策支持以辅助胃癌前疾病诊断。     

Logistic regression and artificial neural network classification models: a methodology review

Logistic regression and artificial neural networks are the models of choice in many medical data classification tasks. In this review, we summarize the differences and similarities of these models from a technical point of view, and compare them with other machine learning algorithms. We provide considerations useful for critically assessing the quality of the models and the results based on these models. Finally, we summarize our findings on how quality criteria for logistic regression and artificial neural network models are met in a sample of papers from the medical literature.     

结合迁移学习与深度卷积网络的心电分类研究

为解决一维深度卷积网络（1D-DCNN）在心电分类方面存在的多类疾病识别不准、难以提取最佳特征等问题,提出一种结合迁移学习与二维深度卷积网络（2D-DCNN）直接识别心电图像的方法。首先,截取R波前后75 ms内的心电信号,并将一维心电电压信号转化为二维灰度图像信号。接着,构建2D-DCNN对心电节拍样本进行分类训练,权值初始化采用在ImageNet大规模图像数据集上进行预训练的AlexNet参数值。本文提出方法在MIT-BIH心电数据库上进行性能验证,其准确率达到98%,并在不同信噪比下保持较高的准确率,证明了所述模型在心电分类上具有良好的鲁棒性。为了验证2D-DCNN的识别性能,实验部分与采用不同激活函数的1D-DCNN、近些年性能较好的深度学习方法进行比较。量化结果表明,结合迁移学习和2D-DCNN方法,比最优1D-DCNN算法,其准确率提升2%、敏感度提升0.6%、特异性提高4%;在二分类与多分类任务中,均好于现有的其他算法。     

Automatic seizure detection using three-dimensional CNN based on multi-channel EEG

Background

Automated seizure detection from clinical EEG data can reduce the diagnosis time and facilitate targeting treatment for epileptic patients. However, current detection approaches mainly rely on limited features manually designed by domain experts, which are inflexible for the detection of a variety of patterns in a large amount of patients’ EEG data. Moreover, conventional machine learning algorithms for seizure detection cannot accommodate multi-channel Electroencephalogram (EEG) data effectively, which contains both temporal and spatial information. Recently, deep learning technology has been widely applied to perform image processing tasks, which could learns useful features from data and process multi-channel data automatically. To provide an effective system for automatic seizure detection, we proposed a new three-dimensional (3D) convolutional neural network (CNN) structure, whose inputs are multi-channel EEG signals.

Methods

EEG data of 13 patients were collected from one center hospital, which has already been inspected by experts. To represent EEG data in CNN, firstly time series of each channel of EEG data was converted into the two-dimensional image. Then all channel images were combined into 3D images according to the mutual correlation intensity between different electrodes. Finally, a CNN was constructed using 3D kernels to predict different stages of EEG data, including inter-ictal, pre-ictal, and ictal stages. The system performance was evaluated and compared with the traditional feature-based classifier and the two-dimensional (2D) deep learning method.

Results

It demonstrated that multi-channel EEG data could provide more information for increasing the specificity and sensitivity in cpmparison result between the single and multi-channel. And the 3D CNN based on multi-channel outperformed the 2D CNN and traditional signal processing methods with an accuracy of more than 90%, an sensitivity of 88.90% and an specificity of 93.78%.

Conclusions

This is the first effort to apply 3D CNN in detecting seizures from EEG. It provides a new way of learning patterns simultaneously from multi-channel EEG signals, and demonstrates that deep neural networks in combination with 3D kernels can establish an effective system for seizure detection.

     

Automated Quality Assessment of Colour Fundus Images for Diabetic Retinopathy Screening in Telemedicine

Fundus images obtained in a telemedicine program are acquired at different sites that are captured by people who have varying levels of experience. These result in a relatively high percentage of images which are later marked as unreadable by graders. Unreadable images require a recapture which is time and cost intensive. An automated method that determines the image quality during acquisition is an effective alternative. To determine the image quality during acquisition, we describe here an automated method for the assessment of image quality in the context of diabetic retinopathy. The method explicitly applies machine learning techniques to access the image and to determine ‘accept’ and ‘reject’ categories. ‘Reject’ category image requires a recapture. A deep convolution neural network is trained to grade the images automatically. A large representative set of 7000 colour fundus images was used for the experiment which was obtained from the EyePACS that were made available by the California Healthcare Foundation. Three retinal image analysis experts were employed to categorise these images into ‘accept’ and ‘reject’ classes based on the precise definition of image quality in the context of DR. The network was trained using 3428 images. The method shows an accuracy of 100% to successfully categorise ‘accept’ and ‘reject’ images, which is about 2% higher than the traditional machine learning method. On a clinical trial, the proposed method shows 97% agreement with human grader. The method can be easily incorporated with the fundus image capturing system in the acquisition centre and can guide the photographer whether a recapture is necessary or not.