期刊
IEEE TRANSACTIONS ON IMAGE PROCESSING
卷 30, 期 -, 页码 472-486出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TIP.2020.3036770
关键词
Image segmentation; Feature extraction; Semantics; Convolution; Task analysis; Principal component analysis; Data mining; Image segmentation; sequence-PCA; attention; model compression; defect detection
资金
- Defense Industrial Technology Development Program [JSZL2019205C003]
- National Natural Science Foundation of China [61971093, 61527803, 61960206010]
- Science and Technology Department of Sichuan, China [2019YJ0208, 2018JY0655, 2018GZ0047]
- Fundamental Research Funds for the Central Universities [ZYGX2019J067]
This article proposes a hybrid multi-dimensional features fusion structure for automated thermography defects detection, with newly designed attention block, Sequence-PCA layer, and lightweight structure that captures semantic information better and improves detection rate. Experimental studies verify the effectiveness and robustness of the proposed model.
This article proposes a hybrid multi-dimensional features fusion structure of spatial and temporal segmentation model for automated thermography defects detection. In addition, the newly designed attention block encourages local interaction among the neighboring pixels to recalibrate the feature maps adaptively. A Sequence-PCA layer is embedded in the network to provide enhanced semantic information. The final model results in a lightweight structure with smaller number of parameters and yet yields uncompromising performance after model compression. The proposed model allows better capture of the semantic information to improve the detection rate in an end-to-end procedure. Compared with current state-of-the-art deep semantic segmentation algorithms, the proposed model presents more accurate and robust results. In addition, the proposed attention module has led to improved performance on two classification tasks compared with other prevalent attention blocks. In order to verify the effectiveness and robustness of the proposed model, experimental studies have been carried out for defects detection on four different datasets. The demo code of the proposed method can be linked soon: http://faculty.uestc.edu.cn/gaobin/zh_CN/lwcg/153392/list/index.htm
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据