Defect Inspection with Tensorflow

A C++ example of defect inspection by running TensorFlow.

DEye 🚀 please find the source code of windows version from the link: https://github.com/sundyCoder/DEye

Any problems please contact with me: Email: [email protected] QQ: 1316501606

0. environment

cuda9.0, cudnn7.3， ubuntu18.04, gcc6.5

1. Install Protobuf

sudo apt-get install autoconf automake libtool curl make unzip
tar -xvf protobuf-all-3.6.0.tar
cd protoc-3.6
./autogen.sh
./configure
make && sudo make install
sudo ldconfig

2.Install Eigen

# Download eigen 3.3.4 from https://github.com/eigenteam/eigen-git-mirror/releases
tar -xvf eigen-eigen-b3f3d4950030.tar.bz2
cd eigen-git-mirror
mkdir build && cd build
cmake ..
make
sudo make install

3.Install TensorFlow

Step 1: 
    vim tensorflow/BUILD
    

462 
463 tf_cc_shared_object(
464     #name = "libtensorflow.so",
465     name = "libCore.so",
466     linkopts = select({
467         "//tensorflow:darwin": [
468             "-Wl,-exported_symbols_list",  # This line must be directly followed by the exported_symbols.lds file
469             "$(location //tensorflow/c:exported_symbols.lds)",
470             "-Wl,-install_name,@rpath/libtensorflow.so",
471         ],
472         "//tensorflow:windows": [],
473         "//conditions:default": [
474             "-z defs",
475             "-Wl,--version-script",  #  This line must be directly followed by the version_script.lds file
476             "$(location //tensorflow/c:version_script.lds)",
477         ],
478     }),
479     visibility = ["//visibility:public"],
             
Step 2: 

rm -fr ~/.cache/bazel*
bazel clean
a. ./configure
b. build
bazel build -c opt --config=monolithic //tensorflow:libCore.so

c. Then Copy the following include headers and dynamic shared library to /usr/local/lib and /usr/local/include:
    mkdir /usr/local/include/tf
    sudo cp -r bazel-genfiles/ /usr/local/include/tf/
    sudo cp -r tensorflow /usr/local/include/tf/
    sudo cp -r third_party /usr/local/include/tf/
    sudo cp -r bazel-bin/libCore.so /usr/local/lib/
    ./tensorflow/contrib/makefile/download_dependencies.sh
    sudo cp -r tensorflow/contrib/makefile/downloads /usr/local/include/tf/tensorflow/contrib/makefile/


d. compile
    g++ -std=c++11 -o tLoader -I/usr/local/include/tf -I/usr/local/include/eigen3 -g -Wall -D_DEBUG -Wshadow -Wno-sign-compare -w  -L/usr/local/lib/libtensorflow_cc `pkg-config --cflags --libs protobuf`  -ltensorflow_cc loader.cpp

4. (Optional)Install/Update cmake

tar xcvf cmake-3.11.1-Linux-x86_64.tar.gz
cd cmake-3.11.1-Linux-x86_64
sudo apt-get purge cmake
sudo cp -r bin /usr/
sudo cp -r share /usr/
sudo cp -r doc /usr/share/
sudo cp -r man /usr/share/

5. install Opencv

sudo apt-get install build-essential checkinstall cmake pkg-config yasm
sudo apt-get install git gfortran
sudo apt-get install libjpeg8-dev libjasper-dev libpng12-dev
sudo apt-get install libtiff5-dev
sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libdc1394-22-dev
sudo apt-get install libxine2-dev libv4l-dev
sudo apt-get install libgstreamer0.10-dev libgstreamer-plugins-base0.10-dev
sudo apt-get install qt5-default libgtk2.0-dev libtbb-dev
sudo apt-get install libatlas-base-dev
sudo apt-get install libfaac-dev libmp3lame-dev libtheora-dev
sudo apt-get install libvorbis-dev libxvidcore-dev
sudo apt-get install libopencore-amrnb-dev libopencore-amrwb-dev
sudo apt-get install x264 v4l-utils
sudo apt-get install python-dev python-pip python3-dev python3-pip
sudo pip3 install numpy scipy matplotlib scikit-image scikit-learn ipython
cd ~
git clone https://github.com/opencv/opencv.git
cd opencv && git checkout 3.4.1
cd ~
git clone https://github.com/opencv/opencv_contrib.git
cd opencv_contrib && git checkout 3.4.1
cd ~/opencv && mkdir build && cd build

cmake -DCMAKE_CXX_FLAGS=-std=c++11 \
    -DCMAKE_BUILD_TYPE=RELEASE \
    -DCMAKE_INSTALL_PREFIX=./install \
    -DBUILD_EXAMPLES=OFF \
    -DBUILD_DOCS=OFF \
    -DBUILD_PERF_TESTS=OFF \
    -DBUILD_TESTS=OFF \
    -DINSTALL_C_EXAMPLES=OFF \
    -DENABLE_PRECOMPILED_HEADERS=OFF \
    -DWITH_OPENMP=ON \
    -DWITH_V4L=OFF \
    -DWITH_TBB=ON \
    -DWITH_QT=OFF \
    -DWITH_OPENGL=ON \
    -DWITH_JPEG=ON \
    -DWITH_FFMPEG=ON \
    -DWITH_GSTREAMER=ON \
    -DWITH_OPENCL=ON \
    -DWITH_GPHOTO2=OFF \
    -DWITH_LIBV4L=OFF \
    -DINSTALL_PYTHON_EXAMPLES=OFF \
    -DBUILD_SHARED_LIBS=ON \
    -DENABLE_CXX11=ON \
    -DOPENCV_EXTRA_MODULES_PATH=../../opencv_contrib/modules ..

    vim ~/.bashrc
    // add the content to .bashrc
    export OpenCV_DIR=/path/to/opencv4.0
    export OpenCV_INCLUDE_DIRS=/path/to/include/opencv4
    export OpenCV_LIBS=/path/to/opencv4.0/lib

6. Testing

a. build libDEye.so
    mkdir build
    cd build
    cmake ..
    make
    cp libDEye.so /usr/local/lib
b. build application
    cd app
    mkdir build
    cd build 
    cmake ..
    make

c. run app    
    cd app/build
./demo

7. model encryption

sh export.sh
cp train_log/pb/frozen_inference_graph.pb TFSecured/python/
python3 encrypt_model.py ./demo/models/frozen_inference_graph.pb ./model.so df6c8cd6696cfe35a6ea8dc14722132420181230

'model.so' is generated.

8. Library

Model file: model.so  model.map
Core library: libDEye.so, libCore.so ( please contact with me if you want)

9.Reference

@article{li2022eid,
  title={EID-GAN: Generative Adversarial Nets for Extremely Imbalanced Data Augmentation},
  author={Li, Wei and Chen, Jinlin and Cao, Jiannong and Ma, Chao and Wang, Jia and Cui, Xiaohui and Chen, Ping},
  journal={IEEE Transactions on Industrial Informatics},
  year={2022},
  publisher={IEEE}
}

https://github.com/lysukhin/tensorflow-object-detection-cpp 
https://tuatini.me/building-tensorflow-as-a-standalone-project/
https://medium.com/@fanzongshaoxing/tensorflow-c-api-to-run-a-object-detection-model-4d5928893b02
https://gist.github.com/kyrs/9adf86366e9e4f04addb
https://medium.com/@fanzongshaoxing/tensorflow-c-api-to-run-a-object-detection-model-4d5928893b02
windows: https://github.com/hluu11/SimpleTF-CPP
https://github.com/sundyCoder/or/tree/master/models/research/tensorrt