DFX SDK Python

libdfxpython is the Python 3 wrapper for the Nuralogix DeepAffex™ SDK

Installation

Download the appropriate wheel file for your OS (Windows, Linux or macOS) and install using pip.

pip install ./libdfx-4.9.0.0-py3-none-linux_x86_64.whl  # Replace filename with the one for your OS

While libdfxpython itself doesn't have any dependencies other than cffi, we highly recommend using opencv-python to interact with it.

Running the example

The example script in this archive (dfxexample.py) requires opencv-python. For a quick test, download the DFX Example archive, which contains:

• Video file - dfxDemoCP1.mov
• Face tracking points - dfxDemoCP1-faces.json

You will also need a study configuration file, which you should have obtained directly from Nuralogix.

In a terminal shell, run:

pip install opencv-python
python dfxexample.py dfxDemoCP1.mov dfxDemoCP1-faces.json study-config-file.dat -o chunks

Example output

The output of running the example is several files (in the chunks folder) which contain the data to be sent to the server. Specifically, the payload0.bin contains the binary payload and properties0.json contains the Server API DataRequest property values describing the payload, for the first chunk.

payload0.bin
properties0.json
metadata0.bin
etc.