search

arturoc / FaceSubstitution

Face substitution experiments using ofxFacetracker
Other
437 stars 120 forks source link

Which versions of Ofx, OfxCV and OfxFacetracker do I need to compile? #29

Closed jBachalo closed 8 years ago

jBachalo commented 8 years ago

Getting lots of compile errors running the latest versions of Ofx, OfxCV and OfxFaceTracker. Does anyone know which versions I need in order to compile FaceSubstitution and other examples?

Another issue is if I need an older version of OSX SDK. El Capitan it seems only works with XCode 7

kylemcdonald commented 8 years ago

ofxCv and ofxFaceTracker are up to date with 0.9.0 but this repo is not maintained. if you want to get it running you have two options:

  1. use OF 0.9.0, create a new project file with the project generator, and fix any bugs that have been introduced with API changes in 0.9.0.
  2. use a version of OF, ofxCv, and ofxFaceTracker that were all released/updated at the same time.
jBachalo commented 8 years ago

Thanks. Option 2 I don't think is an option any more with El Capitan. So will go with option 1.

One question using the new Project Generator In this case all of the FaceSubstituion examples have existing Xcode project files.

Should I use the new 'import' feature in Project generator or create new 'blank' project containing the necessary addons and then copy -overwrite the src files over?

Also does it matter if files are named 'testApp' instead of 'ofApp' ?

Guess I'm not the only one who has attempted getting Facsesubstitution working with latest ofx https://github.com/arturoc/FaceSubstitution/issues/12

Same error

include "Clone.h" in testApp.h

Clone.h file not found.

Copied Clone.h and Clone.cpp over from CloningWrapper project. Then made sure I included the latest FaceTracker addon using the project generator. Success

Now all I need are the 'rules' for good subsititution images. Some work really well, others not at all. Does the pixel resolution, or format matter?

kylemcdonald commented 8 years ago

sounds like you figured it out. as i said on twitter: ambient light is good. avoid shadows, frontal faces, match the resolution of your camera. just experiment :)