Rotoscope: Experiment with dlib, opencv, OSC and animata

Animata is a program that i love, but it has not been maintained for years, it still has a very fast rendering and OSC connection. The OSC connection is used to connect applications to each other.

Sent if I open or close my mouth:

python3 expresion.py | while read A; do
eval "$A"
oscsend localhost 7110 /anibone sf BOCA "$BOCA"
done

Now I sent the horizontal and vertical rotation of my head:

python3 expresion.py | while read A; do
eval "$A"
oscsend localhost 7110 /anibone sf RV "$RV"
oscsend localhost 7110 /anibone sf RH "$RH"
done

Ahora todo junto:

python3 expresion.py | while read A; do
eval "$A"
oscsend localhost 7110 /anibone sf RV "$RV"
oscsend localhost 7110 /anibone sf RH "$RH"
oscsend localhost 7110 /anibone sf BOCA "$BOCA"
done

Now it’s a matter of making the data more fluid and inventing the data that is lost, you should surely use neural networks for that.
I would also like to separate the animate render from the editor and generate animata files in gimp or inkscape. Although I could also use Arap (As-Rigid-As-Possible) for future versions.

Source:

Posted in Experiment, Tutorials and tagged , , , , , .

Leave a Reply

Your email address will not be published. Required fields are marked *

*

code