A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect Detection
In this paper, we propose a highly efficient deep learning-based method for pixel-wise surface defect segmentation algorithm in machine vision. Our method is composed of a segmentation stage (stage 1), a detection stage (stage 2), and a matting stage (stage 3). In the segmentation stage, a lightweig...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8624360/ |
| _version_ | 1819179322891567104 |
|---|---|
| author | Lingteng Qiu Xiaojun Wu Zhiyang Yu |
| author_facet | Lingteng Qiu Xiaojun Wu Zhiyang Yu |
| author_sort | Lingteng Qiu |
| collection | DOAJ |
| description | In this paper, we propose a highly efficient deep learning-based method for pixel-wise surface defect segmentation algorithm in machine vision. Our method is composed of a segmentation stage (stage 1), a detection stage (stage 2), and a matting stage (stage 3). In the segmentation stage, a lightweight fully convolutional network (FCN) is employed to make a pixel-wise prediction of the defect areas. Those predicted defect areas act as the initialization of stage 2, guiding the process of detection to correct the improper segmentation. In the matting stage, a guided filter is utilized to refine the contour of the defect area to reflect the real abnormal region. Besides that, aiming to achieve the tradeoff between efficiency and accuracy, and simultaneously we use depthwise&pointwise convolution layer, strided depthwise convolution layer, and upsample depthwise convolution layer to replace the standard convolution layer, pooling layer, and deconvolution layer, respectively. We validate our findings by analyzing the performance obtained on the dataset of DAGM 2007. |
| first_indexed | 2024-12-22T21:56:37Z |
| format | Article |
| id | doaj.art-f10435f93ccd44f281206b869d2720b5 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-22T21:56:37Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-f10435f93ccd44f281206b869d2720b52022-12-21T18:11:14ZengIEEEIEEE Access2169-35362019-01-017158841589310.1109/ACCESS.2019.28944208624360A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect DetectionLingteng Qiu0Xiaojun Wu1https://orcid.org/0000-0003-4988-5420Zhiyang Yu2School of Mechanical Engineering and Automation, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, ChinaSchool of Mechanical Engineering and Automation, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, ChinaSchool of Mechanical Engineering and Automation, Shenzhen Graduate School, Harbin Institute of Technology, Shenzhen, ChinaIn this paper, we propose a highly efficient deep learning-based method for pixel-wise surface defect segmentation algorithm in machine vision. Our method is composed of a segmentation stage (stage 1), a detection stage (stage 2), and a matting stage (stage 3). In the segmentation stage, a lightweight fully convolutional network (FCN) is employed to make a pixel-wise prediction of the defect areas. Those predicted defect areas act as the initialization of stage 2, guiding the process of detection to correct the improper segmentation. In the matting stage, a guided filter is utilized to refine the contour of the defect area to reflect the real abnormal region. Besides that, aiming to achieve the tradeoff between efficiency and accuracy, and simultaneously we use depthwise&pointwise convolution layer, strided depthwise convolution layer, and upsample depthwise convolution layer to replace the standard convolution layer, pooling layer, and deconvolution layer, respectively. We validate our findings by analyzing the performance obtained on the dataset of DAGM 2007.https://ieeexplore.ieee.org/document/8624360/Depthwise convolutionfully convolutional networkssurface defect segmentationmachine vision |
| spellingShingle | Lingteng Qiu Xiaojun Wu Zhiyang Yu A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect Detection IEEE Access Depthwise convolution fully convolutional networks surface defect segmentation machine vision |
| title | A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect Detection |
| title_full | A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect Detection |
| title_fullStr | A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect Detection |
| title_full_unstemmed | A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect Detection |
| title_short | A High-Efficiency Fully Convolutional Networks for Pixel-Wise Surface Defect Detection |
| title_sort | high efficiency fully convolutional networks for pixel wise surface defect detection |
| topic | Depthwise convolution fully convolutional networks surface defect segmentation machine vision |
| url | https://ieeexplore.ieee.org/document/8624360/ |
| work_keys_str_mv | AT lingtengqiu ahighefficiencyfullyconvolutionalnetworksforpixelwisesurfacedefectdetection AT xiaojunwu ahighefficiencyfullyconvolutionalnetworksforpixelwisesurfacedefectdetection AT zhiyangyu ahighefficiencyfullyconvolutionalnetworksforpixelwisesurfacedefectdetection AT lingtengqiu highefficiencyfullyconvolutionalnetworksforpixelwisesurfacedefectdetection AT xiaojunwu highefficiencyfullyconvolutionalnetworksforpixelwisesurfacedefectdetection AT zhiyangyu highefficiencyfullyconvolutionalnetworksforpixelwisesurfacedefectdetection |