Re: FaceStyle Github Repository

[chrissisura]

Hello Vikas, thank you so much for replying and offering your support! I wanted to build a photobooth for the wedding as a gag and run the software on Jetson Nano. I saw the repository of Ebsynth, but unfortunately it needs the segmentation output of FaceStyle as an input (at least that was my assumption).

[vikasTmz]

As you've mentioned a photo-booth, I'm guessing the camera will capture an image containing primarily faces and upper body parts (something like this), right? Also, how will the users interact with the application i.e are you going to have a default single style or allow the users to have multiple?

I'm asking these questions as this will help you to reduce computation time and figure out which solutions will work best. Also, isn't Jetson Nano primarily for running NN models? Facestyle doesnt use any neural net nor any learning based technique. However, you'll need to use some ML techniques to generate the G_seg.