您的位置: 首页 > 2022年3月 第37卷 第3期 > 文字全文
2023年7月 第38卷 第7期11
目录

人工智能在眼前段疾病的应用

Application of artificial intelligence in anterior segment ophthalmic diseases

来源期刊: 眼科学报 | 2022年3月 第37卷 第3期 171-177 发布时间:2021–07–16 收稿时间:2022/11/28 13:47:39 阅读量:7365
作者:
关键词:
眼科人工智能眼前段计算机辅助诊断系统机器学习
eye artificial intelligence anterior eye segment computer-assisted diagnosis machine learning
DOI:
10.3978/j.issn.1000-4432.2022.03.02
收稿时间:
 
修订日期:
 
接收日期:
 
随着人工智能(artificial intelligence,AI)技术的快速发展,基于深度学习(deep learning,DL)和机器学习的AI技术在医学领域上的应用受到了广泛的关注。AI在眼科的应用也逐渐向更全面更深入的层次发展,通过角膜断层扫描、光学相干断层扫描、裂隙灯图像等技术,AI在对角膜病变、结膜病变、白内障、青光眼等眼部疾病的诊断和治疗方面都表现出了良好的性能。然而AI在眼科的应用方面也存在一些诸如结果可解释性的欠缺、数据集标准化的缺乏、数据集质量的不齐、模型适用性的不足和伦理问题等挑战。在5G和远程医疗飞速发展的时代,眼科AI同时也有许多新的机遇。本文综述了AI在前段眼科疾病中的应用、临床实施的潜在挑战和前景,为AI在眼科领域的进一步发展提供参考信息。
With the rapid development of artificial intelligence (AI) technology, the application of AI technology based on deep learning (DL) and machine learning (ML) in the medical field has received widespread attention. The application of AI in ophthalmology is gradually being shifted to a more comprehensive and in-depth level. Trained on corneal tomography, optical coherence tomography (OCT), slit-lamp images, and other techniques. AI can achieve robust performance in the diagnosis and treatment of corneal lesions, conjunctival lesions, cataract, glaucoma and other ophthalmic diseases. However, there are also some challenges in the application of AI in ophthalmology, including the lack of interpretability of results, lack of standardization of data sets, uneven quality of data sets, insufficient applicability of models and ethical issues. In the era of 5G and telemedicine, there are also many new opportunities for ophthalmic AI. In this review, we provided a summary of the state-of-the-art AI application in anterior segment ophthalmic diseases, potential challenges in clinical implementation and its development prospects, and provides reference information for the further development of artificial intelligence in the field of ophthalmology.
    人工智能(artificial intelligence,AI)这个概念由1956年Mccarthy等在一次学术会议上首次提出,此后AI迅速发展,受到了越来越多人的关注,随后深度学习(deep learning,DL)的出现更加加速了AI的发展[1]。DL是使用人工神经网络结构的机器学习的一个子集[2],它是一种表示学习方法,通常涉及大型的神经网络,包括卷积神经网络(convolutional neural network,CNN)及递归神经网络(recurrent neural network,RNN)等。DL能够自主地从已知的数据中学习,用于对未知数据的预测,因此DL可以用来处理复杂的数据[3]
    在眼科中,AI主要应用于青光眼、白内障、圆锥角膜、年龄相关性黄斑变性、糖尿病性视网膜病变等眼科疾病,其中AI在眼科中最主要的应用是眼后段的疾病,尤其是有关视网膜病变[4],对于眼前段的应用目前还不及眼后段。眼前段疾病由于结构等的复杂性,往往需要专业人士经过多次的检查才能诊断。AI云平台[5]等AI在眼前段疾病的应用均显示出了良好的性能,提高了临床医生诊断的准确性。然而也并不是所有的AI都能够应用于临床,有些AI需要特殊的设备或其测量方法较为复杂使其不能在临床广泛应用。因此本文综述了AI在眼前段疾病的应用,以及现阶段眼科AI所面临的挑战和机遇。

1 AI应用于眼前段疾病

1.1 角膜

    1.1.1 角膜扩张症
    角膜扩张症是一组以局部角膜变薄导致变薄的角膜突出为特征的眼部疾病[6]。准分子激光原位角膜磨镶术(laser in situ keratomileus,LASIK)后医源性角膜扩张是角膜术后最严重的并发症之一[2]。由于这种并发症是不可逆的,并且会损害个体的视觉预后,因此需要AI来帮助临床医生识别具有角膜扩张亚临床特征的患者。
    为了加强对角膜的检测,诸多科研人员研发了许多算法。最初对角膜扩张的研究主要是集中在一个系统上,该系统通过从眼前节分析系统(Orbscan IIz)获得的数据可以检测明显的角膜变化[7]。Orbscan IIz的数据经过Souza等[7]的测试后被证实具有良好的性能,其灵敏度为0.98~1.00,特异度为0.98~1.00。随后,在2013年Smadja等[8-9]的研究证实了采用Scheimpflug成像原理的眼前节分析系统的断层扫描数据优先于Orbscan IIz获得的数据。研究者们已经成功证实ML具有识别显性角膜疾病的能力。Arbelaez等[10]通过单侧圆锥角膜数据进行了一项研究,他们使用机器学习分类器(machine learning classifier,MLC)来区分亚临床圆锥角膜和正常眼睛。在检测亚临床圆锥角膜时,该分类器的精确度为0.973,灵敏度为0.920,特异性为0.977,有良好的性能,可以帮助临床医生诊断亚临床圆锥角膜,但该研究中的亚临床圆锥角膜组是包含具有圆锥角膜扩张症的患者。因此为了进一步提高检测圆锥角膜的准确性,Ambrósio等[2,11]建议使用断层扫描和生物力学指数(tomographic and biomechanical index,TBI),该指数将角膜断层扫描和生物力学分析相结合。该研究回顾性分析了850个圆锥角膜患者,并通过实验数据获得灵敏度为0.904,假阳性率为0.04,有较高的临床价值。Accardo等[12]提出一种基于CNN的圆锥角膜检测算法,该算法通过使用双眼的地形参数提高了神经网络的辨别能力。最终经过临床试验证实了该算法有较好的性能,其灵敏度为94.1%,特异度为97.6%,准确度为96.4%,表明该算法具有筛查圆锥角膜患者的潜力。不同于识别单一角膜疾病的算法,Ruiz Hidalgo等[9-10]开发了一种SVM算法来识别5种不同的角膜模式:圆锥角膜、成形圆锥角膜、散光、屈光手术后和正常。并且通过对于不同角膜的识别,证实了SVM算法的准确率为88.8%,加权平均灵敏度为89.0%,特异度为95.2%,表明该模型在屈光手术前筛查患者的潜力。上述的一些研究对于角膜的识别均有高度的准确性,并且有助于临床医生来识别角膜扩张亚临床特征的患者。
    1.1.2 角膜炎
微生物性角膜炎( microbial keratitis,MK)是全球引起角膜盲的主要眼病之一[13]。但在临床上对角膜炎进行严重程度的分析是具有高度主观性的,严重依赖观察者的诊断,比较耗时耗力[14]。因此,迫切需要一种基于AI的算法来快速准确地诊断微生物性角膜炎。
    Saini等[15]开发了一种自动神经网络,该网络对细菌性和真菌性角膜炎的分类准确率高达90.7%明显高于有经验的人类观察者的预测率(62.8%)。在2018年Wu等[16]利用自适应鲁棒二进制模型(adaptive robust binary pattern,ARBP)结合SVM算法构建了自动诊断算法,高准确度地诊断识别微生物性角膜炎,该方法与角膜刮片结果相比具有良好的优越性。不同于单独识别角膜炎疾病的算法,Wu等[17]开发的深度学习系统能识别角膜炎和其他眼表疾病,识别与诊断相关的信息,并提供治疗建议。虽然目前基于角膜炎的AI的研究尚没有角膜扩张症的多,但是AI对于角膜炎的诊断上面已经展现出了良好的可行性,因此对于角膜炎相关AI的研究有可能作为未来研究的一个方向。

1.2 泪膜

    由于泪膜是由脂质、水和黏蛋白3层组成,当3层中的任何一层出现异常时,泪膜层就会变得不稳定,可能会导致干眼。由于干眼诊断很复杂,没有一个完美的标准[18],目前仍是一项具有挑战性的任务,因此大多数对于干眼的研究均针对干眼的诊断[19-21]
    由脂质层变化引起的泪液蒸发增加是蒸发性干眼(evaporative dry eye,EDE)的特征,利用AI来对泪膜脂质层的厚度和质量进行检测可以有助于对EDE的诊断[22]。脂质层模式评估通过干涉测量法对泪膜表面脂质层进行无创成像,以评估泪膜脂质层的质量和厚度[20]。随后Grus等[23-24]通过人工神经网络对泪膜蛋白质模式的分析,证实了该人工神经网络具有良好的性能,可以作为检测干眼的诊断工具。其准确度为0.93,特异度和敏感度约为90%,这为干眼的准确诊断提供了方向。

1.3 结膜

    结膜充血是眼部炎症常见的体征,是因充血导致结膜不均匀地发红。它常被描述为许多眼病的标志,如各种因素引起的结膜炎、葡萄膜炎和青光眼引起的眼压升高等[25]。目前,对结膜充血严重程度的分级主要依据日本眼部过敏学会(Japan Ocular Allergy Society,JOAS)提出的结膜充血严重程度分级系统[26],但是这种分类却具有一定的主观性[27]。因此,目前尚急需AI来辅助医生诊断结膜充血。
    许多研究人员通过最大似然法来对裂隙灯图像或者普通图像进行评估,消除了专家之间的主观性,同时也减少了计算的时间[28-29]。Yoneda等[27]开发了一种专门用于结膜成像的分析应用程序来建立客观的分级系统。他们发明的充血分析系统是一种通过使用颞球结膜裂隙灯照片的数字处理来量化充血程度的系统。该系统显著地提高了结膜充血严重程度分级的客观性,准确度达70%,但由于测量方法比较复杂,因而目前尚无法普遍使用[27]。目前,一种基于视觉集合群网络(Visual Geometry Group Network)VGG-16深度学习模型的充血严重程度分级系统已被验证为具有高度的准确性,平均加权kappa系数达到了0.74,与临床专家的分级具有高度的一致性[25]。该系统不需要一些特殊的设备和复杂的测量方法,因此较Yoneda等[27]开发的充血分析软件更适合常规应用于临床。Derakhshani等[30]通过评估结膜图像的红绿蓝(red green blue,RGB)层来直接测量结膜血管。该研究按照一定的顺序建立了多层前馈神经网络,其将RGB位图与相对应的人体视觉系统(human visual system,HVS)相匹配,最终通过训练测得人工神经网络的输出与HVS目标的相关系数为0.8865,具有良好的性能。

1.4 白内障

    白内障是指由于晶状体的浑浊而导致晶状体透明度的丧失。白内障是全世界视力障碍的主要原因,占低收入和中等收入国家失明病例的50%以上[31]。因此,白内障的准确诊断和及时手术治疗对于防止视力丧失和提高生活质量至关重要。
    一些深度学习模型应用于白内障诊断的优越性已被证实[32]。AI技术主要通过使用裂隙灯照片和眼底照片来对白内障进行自动诊断和严重程度的分级。对于使用裂隙灯图像的白内障分级诊断中,Li等[33-34]从检测到的晶状体解剖结构中提取一些局部特征建立了一个回归模型,但是该研究并未关注提取特征的可用性和可解释性,导致该模型不能推广到临床。Guo等[35]研究了小波变换和基于草图的方法从眼底图像中提取特征,并应用多类判别分析算法进行白内障检测和分级。多项研究[36-39]表明使用眼底图像来对白内障进行诊断和分级是有优势的:与裂隙灯摄影相比,眼底成像只需要一种成像技术,便于操作和节省时间;眼底成像可同时筛查并诊断其他后段疾病[36-39]。因此,Xiong等[38]通过采用具有多阈值算法的形态学方法,成功地从眼底图像中检测和去除玻璃体混浊,从而将AI算法的精度从77%提高到81%。Xu等[36]开发了一种基于CNN的集成算法,该研究利用8 030幅眼底图像对白内障进行诊断和分级,证实该算法最佳准确性平均能达到86.24%,适合推广到临床。AI云平台CC-Cruiser[40]对于白内障的诊断、治疗和管理也起非常重要的作用。该协作云平台支持各医院之间的患者数据共享,以实现数据集成和患者筛查。当潜在患者来到非专业合作医院进行眼科评估时,他们的人口统计信息、临床数据和联系信息在他们的许可下被收集并立即发送到CC-Cruiser云平台,实现医院之间信息资源的共享。

1.5 青光眼

    青光眼是一组异质性的退行性神经疾病,其特征是视网膜神经节细胞(retinal ganglion ceils,RGCs)及其轴突的进行性丢失[41]。青光眼是一组具有多因素病因的眼部疾病,是全球不可逆失明的主要因素[42-43]。预计2040年青光眼患者人数将达到1.118亿[44]。虽然早期干预可以最大限度地降低青光眼的视力丧失风险,但其无症状性使得在晚期之前难以诊断[45],因此通过AI来对青光眼进行早期的诊断治疗是非常有必要的。
    已有研究人员[46-47]证实了深度学习算法在检测青光眼方面具有良好的性能,他们通过实验测得了DL算法的AUC ≥0.942,比眼科住院医师(0.593~0.640)、主治眼科医师(0.533~0.653)和青光眼专家(0.607~0.663)更准确。对于闭角型青光眼的治疗,光学相干断层扫描显然发挥了重要的作用[48-50]。但是,该方法需要有经验的青光眼专家或训练有素的工作人员来解释光学相干断层扫描的结果,这比较费时费力。在对光学相干断层扫描分析中发现,数字线技术有助于区分青光眼眼睛和正常眼睛,这表明将数字线技术结合到光学相干断层扫描中用于青光眼评估可以潜在地解决当前临床工作中的一些问题[51]。Kucur等[52]研发了一种数字线系统,该系统用视觉诱发电位的视觉表征图像进行训练,最终以87.4%的平均精度识别早期青光眼视觉诱发电位。Xu等[53-54]对前房角区域进行定位,然后直接提取视觉特征进行青光眼亚型的分类。但是,该方法只能报告分类结果,不能提供医生所需的临床测量,因此也不适用于临床。与此同时,Li等[55]开发了一种下行链路网络,该网络主要是在模式偏差图像的概念图上进行训练,目的是为了区分青光眼和非青光眼图像。在评估中该网络实现了87.6%的诊断准确率高于青光眼专家。对于一些高风险的患者也可以使用系列立体视盘照片(stereoscopic photos of optic disc,SPOD),标准自动视野检查(思爱普)或光学相干断层扫描成像进行筛查。一旦确诊,高风险的患者将接受定期临床检查以及SODP、思爱普、眼压和光学相干断层扫描监测[2],降低疾病的发病率。

2 未来的挑战和机遇

    虽然现在AI在眼科当中的应用已经展现出了光明的前景,但现阶段大多数的研究尚处于初步探索阶段,还有许多问题需要解决。1)AI在眼科上的应用需要大量的图像,但是图像在被纳入数据集之前需要专业人员完成标记、注释和质量保证等工作,这会消耗掉大量的人力物力等资源,时间成本也会比较高。虽然迁移学习、无监督学习、弱监督学习等学习方法可以减少对数据的依赖性,但其诊断性能却低于监督学习。因此,目前尚需开发能够运用更少数据、准确性更高的算法,这可能是该领域的一个发展趋势,相信AI在未来会有一个好的发展[56]。2)在大数据时代,不同的中心在进行数据之间交换共享的时候,患者的隐私保护也成为了一个问题。因此,政府应该健全相关的法律法规明确保障患者的隐私。3)数据标准化问题。DL模型是通过训练集来进行学习的,因此训练集的好坏严重影响了最终结果。由于数据集标准化的缺乏,因此建立起来的训练集质量参差不齐,导致最终DL模型的准确性也没有办法保证,没有办法适用于临床。4)虽然已经有许多AI在眼前段应用的研究,但是由于种种原因,这些模型在临床上并不适用,能推广适用于临床的模型较少。因此研发适用于临床的算法模型也是目前所需要的。5)对于结果可解释性的欠缺,即“黑匣子”现象。由于DL是一种端到端的学习方法,即输入原始数据,直接输出结果,因此DL对于检测结果缺乏解释能力,无法对结果提供确切的判断依据。所以,对于临床医生来说,使用DL模型是无法知道完全准确的诊断原因[56]。6)伦理问题:AI能否完全用于眼部疾病的诊断和治疗上面,对于出现的一些错误是需要谁来承担责任?是算法开发人员还是临床医生?一旦在人群中使用错误的算法可能会造成巨大的医源性伤害。因此,在利用AI进行诊断疾病时还需要医生的监督,这也是AI取代不了医生的原因。
    由于疫情的原因,远程医疗得到快速发展。远程医疗可使医疗普及到更为偏远的山村,使更多人得到基本的医疗保障,同时也降低了看病成本,减少了初级保健中眼科医生的压力[57],这也是AI未来发展的一个方向。随着智能手机的普及,智能手机摄影有可能成为未来眼科疾病诊断的一个工具。研究[30,58]已证实一部分AI模型开始应用普通的眼睛照片而不是裂隙灯照片,例如结膜充血分级[30]和检测高眼压[58]等。总之,AI已经融入了我们生活的方方面面,如医疗、生物、语音及驾驶等,相信AI在未来会有一个好的发展。

开放获取声明

    本文适用于知识共享许可协议 (Creative Commons),允许第三方用户按照署名(BY)-非商业性使用(NC)-禁止演绎(ND)(CC BY-NC-ND)的方式共享,即允许第三方对本刊发表的文章进行复制、发行、展览、表演、放映、广播或通过信息网络向公众传播,但在这些过程中必须保留作者署名、仅限于非商业性目的、不得进行演绎创作。详情请访问:https://creativecommons.org/licenses/by-nc-nd/4.0/

1、Lawrence DR, Palacios-González C, Harris J. Artificial Intelligence[J]. Camb Q Healthc Ethics, 2016, 25(2): 250-261.Lawrence DR, Palacios-González C, Harris J. Artificial Intelligence[J]. Camb Q Healthc Ethics, 2016, 25(2): 250-261.
2、Hogarty DT, Mackey DA, Hewitt AW. Current state and future prospects of artificial intelligence in ophthalmology: a review[J]. Clin Exp Ophthalmol, 2019, 47(1): 128-139.Hogarty DT, Mackey DA, Hewitt AW. Current state and future prospects of artificial intelligence in ophthalmology: a review[J]. Clin Exp Ophthalmol, 2019, 47(1): 128-139.
3、LeCun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 2015, 521(7553): 436-444.LeCun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 2015, 521(7553): 436-444.
4、Ting DSW, Peng L, Varadarajan AV, et al. Deep learning in ophthalmology: The technical and clinical considerations[J]. Prog Retin Eye Res, 2019, 72: 100759.Ting DSW, Peng L, Varadarajan AV, et al. Deep learning in ophthalmology: The technical and clinical considerations[J]. Prog Retin Eye Res, 2019, 72: 100759.
5、晏丕松, 项毅帆, 李强, 等. 眼科数据中心和智能服务云平台的建设思路[J]. 眼科学报, 2021, 36(1): 97-103.晏丕松, 项毅帆, 李强, 等. 眼科数据中心和智能服务云平台的建设思路[J]. 眼科学报, 2021, 36(1): 97-103.
6、 Establishment of ophthalmic data center and intelligent service cloud platform[J]. Yan Ke Xue Bao, 2021, 36(1): 97-103. Establishment of ophthalmic data center and intelligent service cloud platform[J]. Yan Ke Xue Bao, 2021, 36(1): 97-103.
7、Romero-Jiménez M, Santodomingo-Rubido J, Wolffsohn JS. Keratoconus: a review[J]. Cont Lens Anterior Eye, 2010, 33(4): 157-166.Romero-Jiménez M, Santodomingo-Rubido J, Wolffsohn JS. Keratoconus: a review[J]. Cont Lens Anterior Eye, 2010, 33(4): 157-166.
8、Souza MB, Medeiros FW, Souza DB, et al. Evaluation of machine learning classifiers in keratoconus detection from orbscan II examinations[J]. Clinics (Sao Paulo), 2010, 65(12): 1223-1228.Souza MB, Medeiros FW, Souza DB, et al. Evaluation of machine learning classifiers in keratoconus detection from orbscan II examinations[J]. Clinics (Sao Paulo), 2010, 65(12): 1223-1228.
9、Smadja D, Touboul D, Cohen A, et al. Detection of subclinical keratoconus using an automated decision tree classification[J]. Am J Ophthalmol, 2013, 156(2): 237-246.Smadja D, Touboul D, Cohen A, et al. Detection of subclinical keratoconus using an automated decision tree classification[J]. Am J Ophthalmol, 2013, 156(2): 237-246.
10、Ruiz Hidalgo I, Rodriguez P, Rozema JJ, et al. Evaluation of a machine-learning classifier for keratoconus detection based on scheimpflug tomography[J]. Cornea, 2016, 35(6): 827-832.Ruiz Hidalgo I, Rodriguez P, Rozema JJ, et al. Evaluation of a machine-learning classifier for keratoconus detection based on scheimpflug tomography[J]. Cornea, 2016, 35(6): 827-832.
11、Arbelaez MC, Versaci F, Vestri G, et al. Use of a support vector machine for keratoconus and subclinical keratoconus detection by topographic and tomographic data[J]. Ophthalmology, 2012, 119(11): 2231-2238.Arbelaez MC, Versaci F, Vestri G, et al. Use of a support vector machine for keratoconus and subclinical keratoconus detection by topographic and tomographic data[J]. Ophthalmology, 2012, 119(11): 2231-2238.
12、Ambrósio R Jr, Lopes BT, Faria-Correia F, et al. Integration of scheimpflug-based corneal tomography and biomechanical assessments for enhancing ectasia detection[J]. J Refract Surg, 2017, 33(7): 434-443.Ambrósio R Jr, Lopes BT, Faria-Correia F, et al. Integration of scheimpflug-based corneal tomography and biomechanical assessments for enhancing ectasia detection[J]. J Refract Surg, 2017, 33(7): 434-443.
13、Accardo PA, Pensiero S. Neural network-based system for early keratoconus detection from corneal topography[J]. J Biomed Inform, 2002, 35(3): 151-159.Accardo PA, Pensiero S. Neural network-based system for early keratoconus detection from corneal topography[J]. J Biomed Inform, 2002, 35(3): 151-159.
14、蓝倩倩, 陈丽妃, 黄慧, 等. 共聚焦显微镜对感染性角膜炎病原学的诊断价值[J]. 中国临床新医学, 2019, 12(6): 626-629.蓝倩倩, 陈丽妃, 黄慧, 等. 共聚焦显微镜对感染性角膜炎病原学的诊断价值[J]. 中国临床新医学, 2019, 12(6): 626-629.
15、 Diagnostic efficiency of confocal microscopy in aetiology of infectious keratitis[J]. Chinese Journal of New Clinical Medicine, 2019, 12(6): 626-629. Diagnostic efficiency of confocal microscopy in aetiology of infectious keratitis[J]. Chinese Journal of New Clinical Medicine, 2019, 12(6): 626-629.
16、Tahvildari M, Singh RB, Saeed HN. Application of artificial intelligence in the diagnosis and management of corneal diseases[J]. Semin Ophthalmol, 2021, 36(8): 641-648.Tahvildari M, Singh RB, Saeed HN. Application of artificial intelligence in the diagnosis and management of corneal diseases[J]. Semin Ophthalmol, 2021, 36(8): 641-648.
17、Saini JS, Jain AK, Kumar S, et al. Neural network approach to classify infective keratitis[J]. Curr Eye Res, 2003, 27(2): 111-116.Saini JS, Jain AK, Kumar S, et al. Neural network approach to classify infective keratitis[J]. Curr Eye Res, 2003, 27(2): 111-116.
18、Wu X, Tao Y, Qiu Q, et al. Application of image recognition-based automatic hyphae detection in fungal keratitis[J]. Australas Phys Eng Sci Med, 2018, 41(1): 95-103.Wu X, Tao Y, Qiu Q, et al. Application of image recognition-based automatic hyphae detection in fungal keratitis[J]. Australas Phys Eng Sci Med, 2018, 41(1): 95-103.
19、Wu X, Qiu Q, Liu Z, et al. Hyphae detection in fungal keratitis images with adaptive robust binary pattern[J]. IEEE Access,2018,6:13449-13460.Wu X, Qiu Q, Liu Z, et al. Hyphae detection in fungal keratitis images with adaptive robust binary pattern[J]. IEEE Access,2018,6:13449-13460.
20、Savini G, Prabhawasat P, Kojima T, et al. The challenge of dry eye diagnosis[J]. Clin Ophthalmol, 2008, 2(1): 31-55.Savini G, Prabhawasat P, Kojima T, et al. The challenge of dry eye diagnosis[J]. Clin Ophthalmol, 2008, 2(1): 31-55.
21、Yedidya T, Hartley R, Guillon JP, et al. Automatic dry eye detection[J]. Med Image Comput Comput Assist Interv, 2007, 10(Pt 1): 792-799.Yedidya T, Hartley R, Guillon JP, et al. Automatic dry eye detection[J]. Med Image Comput Comput Assist Interv, 2007, 10(Pt 1): 792-799.
22、Remeseiro B, Bolon-Canedo V, Peteiro-Barral D, et al. A methodology for improving tear film lipid layer classification[J]. IEEE J Biomed Health Inform, 2014, 18(4): 1485-1493.Remeseiro B, Bolon-Canedo V, Peteiro-Barral D, et al. A methodology for improving tear film lipid layer classification[J]. IEEE J Biomed Health Inform, 2014, 18(4): 1485-1493.
23、Peteiro-Barral D, Remeseiro B, Méndez R, et al. Evaluation of an automatic dry eye test using MCDM methods and rank correlation[J]. Med Biol Eng Comput, 2017, 55(4): 527-536.Peteiro-Barral D, Remeseiro B, Méndez R, et al. Evaluation of an automatic dry eye test using MCDM methods and rank correlation[J]. Med Biol Eng Comput, 2017, 55(4): 527-536.
24、Grus FH, Podust VN, Bruns K, et al. SELDI-TOF-MS ProteinChip array profiling of tears from patients with dry eye[J]. Invest Ophthalmol Vis Sci, 2005, 46(3): 863-876.Grus FH, Podust VN, Bruns K, et al. SELDI-TOF-MS ProteinChip array profiling of tears from patients with dry eye[J]. Invest Ophthalmol Vis Sci, 2005, 46(3): 863-876.
25、Grus FH, Augustin AJ. Analysis of tear protein patterns by a neural network as a diagnostical tool for the detection of dry eyes[J]. Electrophoresis, 1999, 20(4-5): 875-880.Grus FH, Augustin AJ. Analysis of tear protein patterns by a neural network as a diagnostical tool for the detection of dry eyes[J]. Electrophoresis, 1999, 20(4-5): 875-880.
26、Tabuchi H, Masumoto H. Objective evaluation of allergic conjunctival disease (with a focus on the application of artificial intelligence technology)[J]. Allergol Int, 2020, 69(4): 505-509.Tabuchi H, Masumoto H. Objective evaluation of allergic conjunctival disease (with a focus on the application of artificial intelligence technology)[J]. Allergol Int, 2020, 69(4): 505-509.
27、Takamura E, Uchio E, Ebihara N, et al. Japanese guidelines for allergic conjunctival diseases 2017[J]. Allergol Int, 2017, 66(2): 220-229.Takamura E, Uchio E, Ebihara N, et al. Japanese guidelines for allergic conjunctival diseases 2017[J]. Allergol Int, 2017, 66(2): 220-229.
28、Yoneda T, Sumi T, Takahashi A, et al. Automated hyperemia analysis software: reliability and reproducibility in healthy subjects[J]. Jpn J Ophthalmol, 2012, 56(1): 1-7.Yoneda T, Sumi T, Takahashi A, et al. Automated hyperemia analysis software: reliability and reproducibility in healthy subjects[J]. Jpn J Ophthalmol, 2012, 56(1): 1-7.
29、Sánchez Brea ML, Barreira Rodríguez N, Sánchez Maro?o N, et al. On the development of conjunctival hyperemia computer-assisted diagnosis tools: Influence of feature selection and class imbalance in automatic gradings[J]. Artif Intell Med, 2016, 71: 30-42.Sánchez Brea ML, Barreira Rodríguez N, Sánchez Maro?o N, et al. On the development of conjunctival hyperemia computer-assisted diagnosis tools: Influence of feature selection and class imbalance in automatic gradings[J]. Artif Intell Med, 2016, 71: 30-42.
30、Sánchez Brea ML, Barreira Rodríguez N, Mosquera González A, et al. Defining the optimal region of interest for hyperemia grading in the bulbar conjunctiva[J]. Comput Math Methods Med, 2016, 2016: 3695014.Sánchez Brea ML, Barreira Rodríguez N, Mosquera González A, et al. Defining the optimal region of interest for hyperemia grading in the bulbar conjunctiva[J]. Comput Math Methods Med, 2016, 2016: 3695014.
31、Derakhshani R, Saripalle SK, Doynov P. Computational methods for objective assessment of conjunctival vascularity[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2012, 2012: 1490-1493.Derakhshani R, Saripalle SK, Doynov P. Computational methods for objective assessment of conjunctival vascularity[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2012, 2012: 1490-1493.
32、Lee CM, Afshari NA. The global state of cataract blindness[J]. Curr Opin Ophthalmol, 2017, 28(1): 98-103.Lee CM, Afshari NA. The global state of cataract blindness[J]. Curr Opin Ophthalmol, 2017, 28(1): 98-103.
33、Gao X, Lin S, Wong TY. Automatic feature learning to grade nuclear cataracts based on deep learning[J]. IEEE Trans Biomed Eng, 2015, 62(11): 2693-2701.Gao X, Lin S, Wong TY. Automatic feature learning to grade nuclear cataracts based on deep learning[J]. IEEE Trans Biomed Eng, 2015, 62(11): 2693-2701.
34、Li H, Lim JH, Liu J, et al. An automatic diagnosis system of nuclear cataract using slit-lamp images[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2009, 2009: 3693-3696.Li H, Lim JH, Liu J, et al. An automatic diagnosis system of nuclear cataract using slit-lamp images[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2009, 2009: 3693-3696.
35、Li H, Lim JH, Liu J, et al. A computer-aided diagnosis system of nuclear cataract[J]. IEEE Trans Biomed Eng, 2010, 57(7): 1690-1698.Li H, Lim JH, Liu J, et al. A computer-aided diagnosis system of nuclear cataract[J]. IEEE Trans Biomed Eng, 2010, 57(7): 1690-1698.
36、Guo L, Yang J, Peng L, et al. A computer-aided healthcare system for cataract classification and grading based on fundus image analysis[J]. Computers in Industry, 2015, 69: 72-80.Guo L, Yang J, Peng L, et al. A computer-aided healthcare system for cataract classification and grading based on fundus image analysis[J]. Computers in Industry, 2015, 69: 72-80.
37、Xu X, Zhang L, Li J, et al. A hybrid global-local representation CNN model for automatic cataract grading[J]. IEEE J Biomed Health Inform, 2020, 24(2): 556-567.Xu X, Zhang L, Li J, et al. A hybrid global-local representation CNN model for automatic cataract grading[J]. IEEE J Biomed Health Inform, 2020, 24(2): 556-567.
38、Zhang H, Niu K, Xiong Y, et al. Automatic cataract grading methods based on deep learning[J]. Comput Methods Programs Biomed, 2019, 182: 104978.Zhang H, Niu K, Xiong Y, et al. Automatic cataract grading methods based on deep learning[J]. Comput Methods Programs Biomed, 2019, 182: 104978.
39、Xiong L, Li H, Xu L. An approach to evaluate blurriness in retinal images with vitreous opacity for cataract diagnosis[J]. J Healthc Eng, 2017, 2017: 5645498.Xiong L, Li H, Xu L. An approach to evaluate blurriness in retinal images with vitreous opacity for cataract diagnosis[J]. J Healthc Eng, 2017, 2017: 5645498.
40、Yang JJ, Li J, Shen R, et al. Exploiting ensemble learning for automatic cataract detection and grading[J]. Comput Methods Programs Biomed, 2016, 124: 45-57.Yang JJ, Li J, Shen R, et al. Exploiting ensemble learning for automatic cataract detection and grading[J]. Comput Methods Programs Biomed, 2016, 124: 45-57.
41、Lin H, Li R, Liu Z, et al. Diagnostic efficacy and therapeutic decision-making capacity of an artificial intelligence platform for childhood cataracts in eye clinics: a multicentre randomized controlled trial[J]. EClinicalMedicine, 2019, 9: 52-59.Lin H, Li R, Liu Z, et al. Diagnostic efficacy and therapeutic decision-making capacity of an artificial intelligence platform for childhood cataracts in eye clinics: a multicentre randomized controlled trial[J]. EClinicalMedicine, 2019, 9: 52-59.
42、Nicoar? S. The mechanisms of neuronal death in glaucoma[J]. Oftalmologia, 2000, 51(2): 4-6.Nicoar? S. The mechanisms of neuronal death in glaucoma[J]. Oftalmologia, 2000, 51(2): 4-6.
43、Weinreb RN, Aung T, Medeiros FA. The pathophysiology and treatment of glaucoma: a review[J]. JAMA, 2014, 311(18): 1901-1911.Weinreb RN, Aung T, Medeiros FA. The pathophysiology and treatment of glaucoma: a review[J]. JAMA, 2014, 311(18): 1901-1911.
44、Casson RJ, Chidlow G, Wood JP, et al. Definition of glaucoma: clinical and experimental concepts[J]. Clin Exp Ophthalmol, 2012, 40(4): 341-349.Casson RJ, Chidlow G, Wood JP, et al. Definition of glaucoma: clinical and experimental concepts[J]. Clin Exp Ophthalmol, 2012, 40(4): 341-349.
45、Tham YC, Li X, Wong TY, et al. Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis[J]. Ophthalmology, 2014, 121(11): 2081-2090.Tham YC, Li X, Wong TY, et al. Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis[J]. Ophthalmology, 2014, 121(11): 2081-2090.
46、Devalla SK, Liang Z, Pham TH, et al. Glaucoma management in the era of artificial intelligence[J]. Br J Ophthalmol, 2020, 104(3): 301-311.Devalla SK, Liang Z, Pham TH, et al. Glaucoma management in the era of artificial intelligence[J]. Br J Ophthalmol, 2020, 104(3): 301-311.
47、Li Z, He Y, Keel S, et al. Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fundus photographs[J]. Ophthalmology, 2018, 125(8): 1199-1206.Li Z, He Y, Keel S, et al. Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fundus photographs[J]. Ophthalmology, 2018, 125(8): 1199-1206.
48、Ting DSW, Cheung CY, Lim G, et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes[J]. JAMA, 2017, 318(22): 2211-2223.Ting DSW, Cheung CY, Lim G, et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes[J]. JAMA, 2017, 318(22): 2211-2223.
49、Leung CK, Cheung CY, Weinreb RN, et al. Evaluation of retinal nerve fiber layer progression in glaucoma: a comparison between the fast and the regular retinal nerve fiber layer scans[J]. Ophthalmology, 2011, 118(4): 763-767.Leung CK, Cheung CY, Weinreb RN, et al. Evaluation of retinal nerve fiber layer progression in glaucoma: a comparison between the fast and the regular retinal nerve fiber layer scans[J]. Ophthalmology, 2011, 118(4): 763-767.
50、Na JH, Sung KR, Baek S, et al. Detection of glaucoma progression by assessment of segmented macular thickness data obtained using spectral domain optical coherence tomography[J]. Invest Ophthalmol Vis Sci, 2012, 53(7): 3817-3826.Na JH, Sung KR, Baek S, et al. Detection of glaucoma progression by assessment of segmented macular thickness data obtained using spectral domain optical coherence tomography[J]. Invest Ophthalmol Vis Sci, 2012, 53(7): 3817-3826.
51、Na JH, Sung KR, Lee JR, et al. Detection of glaucomatous progression by spectral-domain optical coherence tomography[J]. Ophthalmology, 2013, 120(7): 1388-1395.Na JH, Sung KR, Lee JR, et al. Detection of glaucomatous progression by spectral-domain optical coherence tomography[J]. Ophthalmology, 2013, 120(7): 1388-1395.
52、Ran AR, Tham CC, Chan PP, et al. Deep learning in glaucoma with optical coherence tomography: a review[J]. Eye (Lond), 2021, 35(1): 188-201.Ran AR, Tham CC, Chan PP, et al. Deep learning in glaucoma with optical coherence tomography: a review[J]. Eye (Lond), 2021, 35(1): 188-201.
53、Kucur ?S, Holló G, Sznitman R. A deep learning approach to automatic detection of early glaucoma from visual fields[J]. PLoS One, 2018, 13(11): e0206081.Kucur ?S, Holló G, Sznitman R. A deep learning approach to automatic detection of early glaucoma from visual fields[J]. PLoS One, 2018, 13(11): e0206081.
54、Xu Y, Liu J, Tan NM, et al. Anterior chamber angle classification using multiscale histograms of oriented gradients for glaucoma subtype identification[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2012, 2012: 3167-3170.Xu Y, Liu J, Tan NM, et al. Anterior chamber angle classification using multiscale histograms of oriented gradients for glaucoma subtype identification[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2012, 2012: 3167-3170.
55、Xu Y, Liu J, Cheng J, et al. Automated anterior chamber angle localization and glaucoma type classification in OCT images[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2013, 2013: 7380-7383.Xu Y, Liu J, Cheng J, et al. Automated anterior chamber angle localization and glaucoma type classification in OCT images[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2013, 2013: 7380-7383.
56、Li F, Wang Z, Qu G, et al. Automatic differentiation of glaucoma visual field from non-glaucoma visual filed using deep convolutional neural network[J]. BMC Med Imaging, 2018, 18(1): 35.Li F, Wang Z, Qu G, et al. Automatic differentiation of glaucoma visual field from non-glaucoma visual filed using deep convolutional neural network[J]. BMC Med Imaging, 2018, 18(1): 35.
57、Berner ES, Ozaydin B. Benefits and risks of machine learning decision support systems[J]. JAMA, 2017, 318(23): 2353-2354.Berner ES, Ozaydin B. Benefits and risks of machine learning decision support systems[J]. JAMA, 2017, 318(23): 2353-2354.
58、Parikh D, Armstrong G, Liou V, et al. Advances in telemedicine in ophthalmology[J]. Semin Ophthalmol, 2020, 35(4): 210-215.Parikh D, Armstrong G, Liou V, et al. Advances in telemedicine in ophthalmology[J]. Semin Ophthalmol, 2020, 35(4): 210-215.
59、Aloudat M, Faezipour M, El-Sayed A. High intraocular pressure detection from frontal eye images: a machine learning based approach[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2018, 2018: 5406-5409.Aloudat M, Faezipour M, El-Sayed A. High intraocular pressure detection from frontal eye images: a machine learning based approach[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2018, 2018: 5406-5409.
1、本科教学质量工程项目[教务(2021)93号]。This work was supported by the Undergraduate Teaching Quality Engineering Project, China [(2021) No. 93].()
2、本科教学质量工程项目 [ 教务 (2021)93 号 ]。This work was supported by the Undergraduate Teaching Quality Engineering Project, China [(2021) No. 93]()
上一篇
下一篇
其他期刊
  • 眼科学报

    主管:中华人民共和国教育部
    主办:中山大学
    承办:中山大学中山眼科中心
    主编:林浩添
    主管:中华人民共和国教育部
    主办:中山大学
    浏览
  • Eye Science

    主管:中华人民共和国教育部
    主办:中山大学
    承办:中山大学中山眼科中心
    主编:林浩添
    主管:中华人民共和国教育部
    主办:中山大学
    浏览
推荐阅读
出版者信息
目录