Overview Set up Environment Define the model Prepare the dataset Train the model Quantize the model Convert model to binary Deploy the model on the OpenMV Run smile detection Troubleshooting Next steps
Single Page Download PDF

Set up Environment

Linux machine setup

Set up your Linux environment by installing a virtual machine as described below. For completeness, we illustrate all the steps to follow to set up your own environment.

With Virtual Machine

Note: Installation with Virtual Machine is the recommended option. Installation with Linux will take longer.

  1. Install Virtual Box from this link.
  2. Download the Virtual Box image from this link
  3. Import image. Follow this guide.
  4. Log into machine.
    • Username: Embedded
    • Password: embedded

or

Linux 

If required, follow the steps here.

OpenMV setup

In the host OS (not the Linux virtual machine):

  1. Install the OpenmMV IDE. View the full guide here.
  2. Connect OpenMV to laptop with USB.
  3. Update firmware if necessary (make sure to do this here, it is not possible to do in the virtual machine).

In the VirtualBox, start the OpenMV IDE from the terminal:

./openmvide/bin/openmvide.sh &

Note: If not preinstalled, do the following

  • Install the OpenmMV IDE on the guest OS (the Linux virtual machine). View the full guide here.

Connect board:

  1. Connect OpenMV to laptop with USB.
  2. To setup the USB connection to the OpenMV board on the virtual machine follow this USB VM Setup Guide.

Test:

  1. Click on the connection icon on the bottom left of the IDE or CTRL+E.
  2. Start the program by selecting the play icon on the bottom left of the IDE or CTRL+R.
  3. The program should run and display images on the top right without any errors.

Now that the setup is done, we will deploy the chosen model on the Arm Cortex-M7 OpenMV board.

Read the following document for a full OpenMV guide.

Previous Next