基于LDA的类别约束字典学习的织物疵点图像分类方法
Fabric Defect Image Classification Method Based on LDA Category Constraint Dictionary Learning
DOI: 10.12677/SEA.2022.112024, PDF,    国家自然科学基金支持
作者: 王顺政, 陈影柔, 吕文涛*, 王成群：浙江理工大学信息学院，浙江 杭州；郭 庆：浙江省技术创新服务中心，浙江 杭州；陈亮亮：浙江经贸职业技术学院应用工程系，浙江 杭州
关键词: 字典学习；疵点图像分类；织物疵点特征；线性判别分析；类别约束；Dictionary Learning； Defect Image Classification； Fabric Defect Characteristics； Linear Discriminant Analysis； Category Constraints
摘要: 在基于字典学习的织物疵点图像分类的过程中，由于复杂纹理的织物图像背景结构的复杂性、瑕疵信息的形态多样性、瑕疵信息的隐蔽性，导致基于字典学习的织物疵点分类方法存在不能有效提取织物疵点特征问题，为了解决该问题，本文提出一种针对织物疵点特征的基于LDA (线性判别分析)的类别约束字典学习分类方法。首先，重新构建稀疏表示模型，在抑制类内差异和类间模糊的面向判别性特征的字典学习优势上对稀疏系数进行线性判别约束，自适应获取织物图像的有效信息进行分析，使得字典学习中不同织物图像的稀疏系数具有更好的鉴别能力，获得具有织物判别特征的织物字典。然后，利用织物字典对测试样本进行稀疏表示得到的重构误差向量构建分类器进行分类。最后，在阿里云天池布匹疵点检测数据库、实时工厂收集的织物数据集以及本文课题组采集的少量疵点织物图像上验证本文方法的有效性。在不同的织物数据集实验结果表明，本文提取的织物特征字典分类效果更好。
Abstract: In the process of fabric defect image classification based on dictionary learning, due to the complex-ity of the background structure of complex texture fabric image, the morphological diversity of de-fect information and the concealment of defect information, the fabric defect classification method based on dictionary learning cannot effectively extract the characteristics of fabric defects. In order to solve this problem, this paper proposes a category constrained dictionary learning classification method based on LDA (linear discriminant analysis) for the characteristics of fabric defects. Firstly, the sparse representation model is reconstructed. Based on the discrimination oriented dictionary learning advantage of suppressing intra class differences and inter class fuzziness, the sparse coeffi-cients are linearly discriminated and constrained, and the effective information of fabric images is obtained and analyzed adaptively, so that the sparse coefficients of different fabric images in dic-tionary learning have better discrimination ability. A fabric dictionary with fabric discrimination characteristics is obtained. Then, the fabric dictionary is used to sparse representation of the test samples, and the reconstructed error vector is used to construct a classifier for classification. Finally, the effectiveness of this method is verified on Alibaba cloud Tianchi fabric defect detection database, fabric data sets collected by real-time factories and a small number of defective fabric images col-lected by our research group. The experimental results on different fabric data sets show that the fabric feature dictionary extracted in this paper has better classification effect.
文章引用：王顺政, 陈影柔, 吕文涛, 郭庆, 陈亮亮, 王成群. 基于LDA的类别约束字典学习的织物疵点图像分类方法[J]. 软件工程与应用, 2022, 11(2): 223-234. https://doi.org/10.12677/SEA.2022.112024

参考文献

[1] 赵艳, 左保齐. 机器视觉在织物疵点检测上的应用研究综述[J]. 计算机工程与应用, 2020, 56(2): 11-17.
[2] 吕文涛, 林琪琪, 钟佳莹, 王成群, 徐伟强. 面向织物疵点检测的图像处理技术研究进展[J]. 纺织学报, 2021, 42(11): 197-206.
[3] 邹超, 汪秉文, 孙志刚. 基于机器视觉的织物疵点检测方法综述[J]. 天津工业大学学报, 2009, 28(2): 78-82+85.
[4] 彭涛, 彭迪, 刘军平, 胡新荣, 张自力, 陈常念, 等. 基于图卷积神经网络的织物分类研究[J]. 计算机应用研究, 2021, 38(5): 1581-1585+1594.
[5] 尹光灿, 罗戎蕾. 基于卷积神经网络的服装领型识别与分类研究[J]. 现代纺织技术, 2020, 28(3): 48-53.
[6] Jun, X., Wang, J.G., Zhou, J., Meng, S., Pan, R. and Gao, W. (2020) Fabric Defect Detection Based on a Deep Convolutional Neural Network Using a Two-Stage Strategy. Textile Research Journal, 91, 130-142. [Google Scholar] [CrossRef] 
[7] Wei, B., Hao, K.R., Tang, X.S. and Ding, Y. (2019) A New Method Using the Convolutional Neural Network with Compressive Sensing for Fabric Defect Classification Based on Small Sample Sizes. Textile Research Journal, 89, 3539-3555. [Google Scholar] [CrossRef] 
[8] 练秋生, 石保顺, 陈书贞. 字典学习模型、算法及其应用研究进展[J]. 自动化学报, 2015, 41(2): 240-260.
[9] Zhan, Z., Li, L.Q., Chen, X. and Wang, J. (2020) Discriminative-Shared Dictionary Learning for Class-Specific Fabric Texture Characterization. Textile Research Journal, 90, 2478-2491. [Google Scholar] [CrossRef] 
[10] Xu, J., Xiang, L., Liu, Q.S., Gilmore, H., Wu, J., Tang, J., et al. (2016) Stacked Sparse Autoencoder (SSAE) for Nuclei Detec-tion on Breast Cancer Histopathology Images. IEEE Transactions on Medical Imaging, 35, 119-130. [Google Scholar] [CrossRef] 
[11] 崔振, 山世光, 陈熙霖. 结构化稀疏线性判别分析[J]. 计算机研究与发展, 2014, 51(10): 2295-2301.
[12] 陈子鎏, 胡高鹏, 王晓明, 黄增喜, 杜亚军. 基于局部类内结构的鉴别性字典学习方法[J]. 计算机应用研究, 2021, 38(2): 489-494.
[13] 吴莹, 占竹, 汪军. 基于通用字典的局部机织物纹理稳定表征[J]. 纺织学报, 2019, 40(11): 45-49+56.
[14] Yang, M., Zhang, L., Feng, X.C. and Zhang, D. (2014) Sparse Representation Based Fisher Discrimination Dictionary Learning for Image Classification. International Journal of Computer Vision, 109, 209-232. [Google Scholar] [CrossRef] 
[15] Jiang, Z.L., Lin, Z. and Davis, L.S. (2013) Label Consistent K-SVD: Learning a Discriminative Dictionary for Recognition. IEEE Transactions on Pattern Analysis and Machine In-telligence, 35, 2651-2664. [Google Scholar] [CrossRef] 
[16] Vu, T.H., Mousavi, H.S., Monga, V., Rao, G. and Rao, U.K.A. (2016) Histopathological Image Classification Using Discriminative Feature-Oriented Dictionary Learning. IEEE Trans-actions on Medical Imaging, 35, 738-751. [Google Scholar] [CrossRef] 
[17] Martinez, A.M., Zhu, M.L. (2005) Where Are Linear Feature Ex-traction Methods Applicable? IEEE Transactions on Pattern Analysis and Machine Intelligence, 27, 1934-1943. [Google Scholar] [CrossRef] 
[18] Wen, J., Fang, X.Z., Cui, J.R., Fei, L., Yan, K., Chen, Y., et al. (2019) Robust Sparse Linear Discriminant Analysis. IEEE Transactions on Circuits and Systems for Video Technology, 29, 390-403. [Google Scholar] [CrossRef] 
[19] Mairal, J., Bach, F., Ponce, J. and Sapiro, G. (2009) Online Learning for Matrix Factorization and Sparse Coding. Journal of Machine Learning Research, 11, 19-60.
[20] Tropp, J.A. and Gilbert, A.C. (2007) Signal Recovery from Random Measurements via Orthogonal Matching Pursuit. IEEE Transactions on Information Theory, 53, 4655-4666. [Google Scholar] [CrossRef] 
[21] Wright, J., Yang, A.Y., Ganesh, A., Sastry, S.S. and Ma, Y. (2009) Robust Face Recognition via Sparse Representation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 31, 210-227. [Google Scholar] [CrossRef] 

为你推荐