Introduction

This directory includes the SSD MobileNet network and example code. It is a caffe based object recognition and classification network. Please take a look at this github repository for more information: https://github.com/chuanqi305/MobileNet-SSD

The provided Makefile does the following:

1. Downloads network prototxt and caffemodel.

2. Compiles the network files into a graph file for the NCS device.

3. Runs the provided run.py program which creates a GUI window that shows the results of the network for a single image.

Prerequisites

This program requires:

• 1 NCS device

• NCSDK 1.11 or greater

• run 'make install' and then 'make examples'

• OpenCV

Makefile

Provided Makefile has various targets that help with the above mentioned tasks.

make help

Shows available targets.

make all

Builds and/or gathers all the required files needed to run the application except the ncsdk. This must be done as a separate step.

make run_py

Runs the provided python program which runs a single inference on a single image and displays the output in a GUI window as well as on the console.

make clean

Removes all the temporary files that are created by the Makefile

Note

Currently the caffemodel file is not downloaded via the make file. Instead the compiled graph file is included in the repository. This is because the owner of the model file has hosted it on google drive which isn't easily downloaed via wget in the make file. A message will print to the console rather than downloading the model and/or compiling it to a graph file.