Autolabeling-Enhanced Active Learning for Cost-Efficient Surface Defect Visual Classification

Autolabeling-Enhanced Active Learning for Cost-Efficient Surface Defect Visual Classification
复制标题

自动标记增强型主动学习可实现经济高效的表面缺陷视觉分类

DOI:
10.1109/tim.2020.3032190
复制
发表时间:
2021
影响因子:
5.6
通讯作者:
Yin Zhouping
Yin Zhouping
中科院分区:
工程技术2区
文献类型:
--
作者:
Yang Hua;Song Kaiyou;Mao Fangqin;Yin Zhouping

文献摘要

参考文献

相似文献

主动学习可以减少标记训练样本所需的人力,同时保持视觉分类器的性能。然而,现有的主动学习框架不能用于对工业产品表面缺陷进行视觉分类,因为它们仍然需要大量的手动注释工作。在这项研究中,我们提出了一种经济高效的自动标记增强主动学习(ALEAL)框架,以减少表面缺陷视觉分类所需的人工注释工作。所提出的 ALEAL 框架采用深度卷积神经网络 (CNN) 作为视觉分类器,根据一组初始的人类标记训练样本进行训练。然后,将收集到的未标记训练样本输入到分类器中进行类别置信度估计。接下来,提出了一种新颖的多样化成本有效查询策略(DCEQS)来选择一些高置信度样本进行自动标记,并选择一些信息丰富的样本作为需要标记的样本提案。随后,为了进一步减少人工注释工作,ALEAL 中提出并引入了一种新颖的自动标记模块,该模块可以自动标记 DCEQS 选择的信息丰富的未标记训练样本的一部分。在本研究中,通过测量未标记和标记样本之间的相似性,提出了一种新颖的基于注意力的相似性测量网络(ASMN)作为该自动标记模块的实现。最后,剩余的未标记样本由人类专家进行注释,所有新标记的样本都用于重新训练分类器。通过 DCEQS 和 ASMN 的自动标记过程,ALEAL 可以自动标记额外的训练样本,并在需要很少的人工标记训练样本的情况下实现有竞争力的性能,这在工业应用中非常重要。大量实验结果表明,与流行的主动学习方法相比,ALEAL 可以显着减少人工注释的工作量,并在工业产品表面缺陷的视觉分类方面实现最先进的成本效率。
Active learning can reduce the human effort required for labeling training samples while preserving the performance of visual classifiers. However, existing active learning frameworks cannot be used to perform visual classification of industrial product surface defects because they still require intensive manual annotation efforts. In this study, we propose a cost-efficient autolabeling-enhanced active learning (ALEAL) framework to reduce the human annotation effort required for surface defect visual classification. The proposed ALEAL framework employs a deep convolutional neural network (CNN) as a visual classifier trained from an initial set of human-labeled training samples. Then, the collected unlabeled training samples are input into the classifier for category confidence estimation. Next, a novel diverse cost-effective query strategy (DCEQS) is proposed to select some high-confidence samples for autolabeling and some informative samples for sample proposals that need labeling. Subsequently, to further reduce the human annotation effort, a novel autolabeling module is proposed and introduced in ALEAL that can automatically label a portion of the informative unlabeled training samples selected by the DCEQS. In this study, a novel attention-based similarity measurement network (ASMN) is proposed as an implementation of this autolabeling module by measuring the similarity between unlabeled and labeled samples. Finally, the remaining unlabeled samples are annotated by human experts, and all the newly labeled samples are used to retrain the classifier. Through the autolabeling process from the DCEQS and ASMN, ALEAL can automatically label additional training samples and achieve a competitive performance while requiring few human-labeled training samples, which is highly important in industrial applications. Extensive experimental results show that, compared with popular active learning methods, ALEAL can dramatically reduce the effort involved in human annotation and achieve state-of-the-art cost efficiency for the visual classification of industrial product surface defects..
DOI: --
发表时间: 2004
期刊: --
影响因子: --
作者:
Mario Fritz;E. Hayman;B. Caputo;J. Eklundh
通讯作者: Mario Fritz;E. Hayman;B. Caputo;J. Eklundh
DOI: 10.21236/ada440382
发表时间: 2005-07
期刊: Thin Solid Films
影响因子: 2.1
作者:
A. Culotta;A. McCallum
通讯作者: A. Culotta;A. McCallum
DOI: --
发表时间: 2003-08
期刊: --
影响因子: --
作者:
K. Brinker
通讯作者: K. Brinker
DOI: 10.1109/tsm.2015.2430453
发表时间: 2015-05
影响因子: 2.7
作者:
Buyang Zhang;Hua Yang;Z. Yin
通讯作者: Buyang Zhang;Hua Yang;Z. Yin
DOI: 10.1002/j.1538-7305.1948.tb01338.x
发表时间: 1948-01-01
影响因子: --
作者:
SHANNON, CE
通讯作者: SHANNON, CE