文献类型：期刊文献

英文题名：Multiview Wasserstein generative adversarial network for imbalanced pearl classification

作者：Gao, Shuang[1];Dai, Yun[2];Li, Yingjie[3];Liu, Kaixin[2];Chen, Kun[1];Liu, Yi[2]

机构：[1]Shaoxing Univ, Sch Mech & Elect Engn, Shaoxing 312000, Peoples R China;[2]Zhejiang Univ Technol, Inst Proc Equipment & Control Engn, Hangzhou 310023, Peoples R China;[3]China Mobile Commun Grp Zhejiang Co Ltd, Huzhou Branch, Huzhou 313000, Peoples R China

年份：2022

卷号：33

期号：8

外文期刊名：MEASUREMENT SCIENCE AND TECHNOLOGY

收录：SCI-EXPANDED(收录号：WOS:000793497300001)、、EI(收录号：20222112145251)、Scopus(收录号：2-s2.0-85130447492)、WOS

基金：This research was funded by the National Natural Science Foundation of China (Grant Nos. 62022073 and 61873241) and the Natural Science Foundation of Zhejiang Province (Grant No. LQ22E090007) for their financial support.

语种：英文

外文关键词：imbalanced learning; pearl classification; generative adversarial network; deep learning; convolutional neural network

外文摘要：This work described in this paper aims to enhance the level of automation of industrial pearl classification through deep learning methods. To better extract the features of different classes and improve classification accuracy, balanced training datasets are usually needed for machine learning methods. However, the pearl datasets obtained in practice are often imbalanced; in particular, the acquisition cost of some classes is high. An enhanced generative adversarial network, named the multiview Wasserstein generative adversarial network (MVWGAN), is proposed for the imbalanced pearl classification problem. For the minority classes in the training datasets, the MVWGAN method can generate high-quality multiview images simultaneously to balance the original imbalanced datasets. The augmented balanced datasets are used to train a multistream convolution neural network (MS-CNN) for pearl classification. The experimental results show that MVWGAN can overcome the imbalanced learning problem and improve the classification performance of MS-CNN effectively. Moreover, feature visualization is implemented to intuitively explain the effectiveness of MVWGAN.