Skip to content

FanZhang91/DEye

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DEye (Keep an Eye on Defects Inspection)


1. Abstract

Defect Eye is an open source software library based on tensorflow1.4, which focus on surface defect inspection. The application area cover the full range of yield applications within the manufacturing environment, including incoming process tool qualification, wafer qualification, glass surface qualification, reticle qualification, research and development, and tool, process and line monitoring. Patterned and unpatterned wafer defect inspection and qualification tools find particles and pattern defects on the front surface, back surface and edge of the wafer, allowing engineers to detect and monitor critical yield excursions. Also, It can be used for medical image inpsection, including Lung PET/CT,breast MRI, CT Colongraphy, Digital Chest X-ray images.

DEye

2. Usage

Compiled tensorflow-r1.4 GPU version using CMake,VisualStudio 2015, CUDA8.0, cudnn6.0.

How to use DEye

  • Install VisualStudio Community2015 Install NVIDIA CUDA 8.0

  • git clone https://github.com/sundyCoder/DEye

  • Download tensorflow.dll, place it under DEye/bin

  • Download tensorflow.lib and libprotobuf.lib, place theme under DEye/extra/tensorflow-r1.4/

  • Download inception_v3_2016_08_28_frozen.pb, place it under DEye/data

  • Open Visual Studio Solution "DEye.sln" which should be under DEye/build/vc14,Solution configurations option choose "Release", Soluton Platform option choose "x64".

  • Build and run the GUI project, you can do model training for your inspection cases.

3. Applications

3.1 IC Chips Defects Inspection

3.2 Highway Road Crack Damage Inpection

3.3 Fabric Defects Inpection

3.4 Cover Glass Inpection

3.5 Civil Infrastructure Defect Detection

3.6 Power lines Crack Detection

3.7 Medical Image Classification

4. Datasets

  1. Weakly Supervised Learning for Industrial Optical Inspection

    DAGM: https://hci.iwr.uni-heidelberg.de/node/3616

  2. Micro surface defect database

    https://pan.baidu.com/s/1QM0AxlGjUlkHHyxwamIMmA

  3. Oil pollution defect database

    https://pan.baidu.com/s/1_aU_Bfh7lcxpYW1no2MlUQ

  4. Bridge Crack Image Data

    http://pan.baidu.com/s/1bplPrPl

  5. ETHZ Datasets

    ETHZ: http://www.vision.ee.ethz.ch/en/datasets/

  6. RSDDs dataset

    http://icn.bjtu.edu.cn/Visint/resources/RSDDs.aspx

  7. Crack Forest Datasets

    https://github.com/cuilimeng/CrackForest

  8. CV Datasets on the Web

    http://www.cvpapers.com/datasets.html

  9. An RGB-D dataset and evaluation methodology for detection and 6D pose estimation of texture-less objects

    http://cmp.felk.cvut.cz/t-less/

  10. ** CV Dataset on the web **

    http://cvpapers.com/datasets.html

5. Contact

Email: [email protected]
QQ:    1316501606

Welcome to support money for future open source development. (total: 231.66 RMB)

Notice: Any comments and suggetions are welcomed, kindly please introduce yourself(name, country, organization etc.) when contact with me, thanks for your cooperation.

6. License

Apache License 2.0

7. Citation

Use this bibtex to cite this repository:

@misc{DEye,
  title={CNN-based Manufacturing Defect Detection on Tensorflow},
  author={Sundy},
  year={2017},
  publisher={Github},
  journal={GitHub repository},
  howpublished={\url{https://github.com/sundyCoder/DEye}},
}

About

Keep an Eye on Defects Inspection.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 97.8%
  • Python 1.8%
  • C 0.2%
  • CMake 0.1%
  • C# 0.1%
  • Shell 0.0%