Anomaly Composition and Decomposition Network for Accurate Visual Inspection of Texture Defects
Anomaly Composition and Decomposition Network for Accurate Visual Inspection of Texture Defects
复制标题
用于纹理缺陷精确视觉检查的异常合成和分解网络
DOI：
10.1109/tim.2022.3196133
复制
发表时间：
2022
影响因子：
5.6
通讯作者：
Zhouping Yin
中科院分区：
文献类型：
--
作者：
Kaiyou Song;Hua Yang;Zhouping Yin
由于各种纹理和缺陷的极端变化，纹理缺陷检测仍然具有挑战性。目前基于无监督学习的纹理缺陷检测方法由于缺乏一种明确的机制来激励模型为缺陷创建较大的异常分数，因此无法同时检测多种纹理缺陷。在这项研究中，我们提出了一种新的异常组成和分解网络（ACDN）来精确检测各种纹理缺陷。在ACDN中，提出了一种基于高斯采样的异常合成（GSAC）方法来执行异常合成过程，该方法将大量缺陷图像合成用于训练。然后，提出了一种新的异常分解网络（ADN）来执行异常分解过程，该网络通过迫使异常图像的固有纹理特征与无缺陷图像共享一个共同的分布，将缺陷图像分解为纹理背景图像和异常图像。通过GSAC和ADN， ACDN不仅可以准确地重建纹理背景图像，对缺陷区域造成较大的重建误差，而且可以准确地分割缺陷。在测试阶段，通过训练好的ADN将缺陷图像分解为纹理背景图像和异常图像。然后将缺陷图像与纹理背景图像之间的残差图像与异常图像融合，得到缺陷检测结果。在主流纹理缺陷数据集上的大量实验结果表明，ACDN能够达到最先进的纹理缺陷检测精度。
Texture defect inspection remains challenging due to the extreme variations in various textures and defects. Current unsupervised learning-based texture defect inspection methods cannot simultaneously inspect a wide variety of texture defects because they lack an explicit mechanism to encourage the model to create large anomaly scores for defects. In this study, we propose a novel anomaly composition and decomposition network (ACDN) for accurate inspection of various texture defects. In the proposed ACDN, a Gaussian-sampling-based anomaly composition (GSAC) method is proposed to perform the anomaly composition procedure, which composites a large number of defective images for training. Then, a novel anomaly decomposition network (ADN) is proposed to perform the anomaly decomposition procedure, which decomposes the defective images into texture background images and anomaly images by forcing the intrinsic texture features of abnormal images to share a common distribution with those of defect-free images. Through the GSAC and ADN, ACDN learns not only to accurately reconstruct texture background images to cause large reconstruction errors for defect regions but also accurately segment defects. In the testing phase, a defective image is decomposed into a texture background image and an anomaly image through the trained ADN. The residual image between the defective image and the texture background image is then fused with the anomaly image to obtain the defect inspection result. Extensive experimental results on mainstream texture defect datasets demonstrate that ACDN achieves state-of-the-art texture defect inspection accuracy.
登录
查看更多内容
影响因子：
7.5
作者：
Tax, DMJ;Duin, RPW
通讯作者：
Duin, RPW
DOI：
10.1109/tim.2020.3038413
发表时间：
2020-11
影响因子：
5.6
作者：
Chengkan Lv;Fei Shen;Zhengtao Zhang;De Xu;Yonghao He
通讯作者：
Chengkan Lv;Fei Shen;Zhengtao Zhang;De Xu;Yonghao He
影响因子：
19.5
作者：
Russakovsky, Olga;Deng, Jia;Fei-Fei, Li
通讯作者：
Fei-Fei, Li
从 BoW 到 CNN：纹理分类的纹理表示的两个十年
影响因子：
19.5
作者：
Liu, Li;Chen, Jie;Pietikainen, Matti
通讯作者：
Pietikainen, Matti
DOI：
--
发表时间：
2006
期刊：
--
影响因子：
--
作者：
Geoffrey E. Hinton;R. Salakhutdinov
通讯作者：
Geoffrey E. Hinton;R. Salakhutdinov