RZ/G Shopping Basket Demo Application Source Code

This repository contains the code required to build the application demo. This demo requires TensorFlow and OpenCV.

Manual Build Instructions

RZ/G2

1. Setup the yocto enviroment described in meta-renesas-ai-demos (copy .conf files from templates) and run bitbake core-image-qt-sdk -c populate_sdk to create the cross toolchain.
2. Install cross toolchain with sudo sh ./poky-glibc-x86_64-core-image-qt-sdk-aarch64-toolchain-<SDK Version>.sh.
3. Set up environment variables with source /<SDK location>/environment-setup-aarch64-poky-linux.
4. Run qmake.
5. Run make.
6. Copy shoppingbasket_demo_app to the root filesystem.
7. Copy shoppingBasketDemo.tflite to /opt/shopping-basket-demo.
8. Run the app with ./shoppingbasket_demo_app.

Ubuntu

1. Install dependencies

   sudo apt install cmake qtbase5-dev qtdeclarative5-dev qt5-default qtmultimedia5-dev qtcreator
   

2. Install opencv core and opencv videoio, make sure your version has Gstreamer enabled. Otherwise build and install OpenCV

   git clone  https://github.com/opencv/opencv.git
   cd opencv/
   mkdir build/
   cd build/
   cmake -D CMAKE_BUILD_TYPE=RELEASE \
   -D CMAKE_INSTALL_PREFIX=/usr \
   -D WITH_V4L=ON \
   -D WITH_QT=ON \
   -D WITH_GSTREAMER=ON ..
   make -j$(nproc)
   sudo make -j$(nproc) install
   

3. Build and install TensorFlow lite v2.3.1

   git clone  https://github.com/tensorflow/tensorflow.git
   cd tensorflow/
   git checkout v2.3.1
   ./tensorflow/lite/tools/make/download_dependencies.sh
   make -j$(nproc) -f ./tensorflow/lite/tools/make/Makefile
   sudo cp ./tensorflow/lite/tools/make/gen/linux_x86_64/lib/libtensorflow-lite.a /usr/local/lib/
   sudo cp -r tensorflow/ /usr/local/include
   sudo cp -r tensorflow/lite/tools/make/downloads/flatbuffers/include/flatbuffers /usr/local/include
   

4. Copy shoppingBasketDemo.tflite to /opt/shopping-basket-demo

   sudo mkdir /opt/shopping-basket-demo
   cd /opt/shopping-basket-demo
   sudo wget https://github.com/renesas-rz/meta-renesas-ai-demos/raw/master/meta-shopping-basket-demo/recipes-ai/shopping-basket-demo/files/shoppingBasketDemo.tflite
   

5. Exclude ArmNN incompatible code

   As ArmNN is not supported for X86 machines, remove the ArmNN code by uncommenting the line below from shoppingbasket_demo_app.pro:

   #DEFINES += SBD_X86
   

6. Build demo application

   cd rzg-shopping-basket-demo
   qmake
   make -j$(nproc)
   sudo cp shoppingbasket_demo_app /opt/shopping-basket-demo
   

7. Run the demo with /opt/shopping-basket-demo/shoppingbasket_demo_app