详情页 - 首都医科大学附属北京同仁医院知识库

当前位置：首页 > 详情页

A fundus image classification framework for learning with noisy labels

1| 认领 | 导出 | 链接全文 |

文献详情

资源类型：

WOS体系：

Pubmed体系：

收录情况： ◇ SCIE

作者：

机构： [1]Beijing Institute of Technology, No. 5, Zhong Guan Cun South Street, Beijing, 100081, China [2]Beijing Institute of Ophthalmology, Beijing Tongren Hospital, Capital Medical University, Beijing, 100730, China

出处：

DOI：

ISSN：

关键词： Fundus diseases classification Noisy labels Confidence learning Negative learning Sharpness-aware minimization

摘要：

Fundus images are widely used in the screening and diagnosis of eye diseases. Current classification algorithms for computer-aided diagnosis in fundus images rely on large amounts of data with reliable labels. However, the appearance of noisy labels degrades the performance of data-dependent algorithms, such as supervised deep learning. A noisy label learning framework suitable for the multiclass classification of fundus diseases is presented in this paper, which combines data cleansing (DC), adaptive negative learning (ANL), and sharpness-aware minimization (SAM) modules. Firstly, the DC module filters the noisy labels in the training dataset based on the prediction confidence. Then, the ANL module modifies the loss function by choosing complementary labels, which are neither the given labels nor the labels with the highest confidence. Moreover, for better generalization, the SAM module is applied by simultaneously optimizing the loss and its sharpness. Extensive experiments on both private and public datasets show that our method greatly promotes the performance for classification of multiple fundus diseases with noisy labels.Copyright © 2023 Elsevier Ltd. All rights reserved.

基金：

语种：

被引次数：

WOS：

PubmedID：

中科院(CAS)分区：

出版当年[2022]版：

大类 | 2 区工程技术

小类 | 2 区工程：生物医学 2 区核医学

最新[2023]版：

大类 | 2 区医学

小类 | 2 区工程：生物医学 2 区核医学

JCR分区：

出版当年[2021]版：

Q1 ENGINEERING, BIOMEDICAL Q1 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

最新[2023]版：

Q1 ENGINEERING, BIOMEDICAL Q1 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING

影响因子： 5.4 最新[2023版] 6.1 最新五年平均 7.422 出版当年[2021版] 6.141 出版当年五年平均 4.79 出版前一年[2020版] 5.7 出版后一年[2022版]

第一作者：

第一作者机构： [1]Beijing Institute of Technology, No. 5, Zhong Guan Cun South Street, Beijing, 100081, China

通讯作者：

推荐引用方式(GB/T 7714)：

APA：

MLA：

相关文献

[1]An auxiliary diagnostic tool for common fundus diseases based on fundus color photography and light-weight classification models [2]Automatic Multilabel Classification of Multiple Fundus Diseases Based on Convolutional Neural Network With Squeeze-and-Excitation Attention [3]Glaucomatous Injury of Central Nerve System [4]患高血压冬天护好三部位 [5]前列腺等离子剜切术和钬激光剜除术治疗良性前列腺增生疗效分析 [6]新型冠状病毒肺炎放射诊断检查中感染控制与放射卫生防护管理 [7]萎缩型老年性黄斑变性患眼地图样萎缩面积的多模态成像定量分析 [8]苯扎托品对大鼠非动脉炎性前部缺血性视神经病变模型神经保护作用初探 [9]甲状腺肿瘤的常见手术方式及icd-9-cm-3编码解析 [10]Observation of fundus and optical coherence tomography angiography manifestations in carotid cavernous fistula patients