Works in preview mode. Must add binary model files to C3 project files. See example.
Integrate face-api.js as a C3 plugin. Minimal experimental implementation for testing. Currently uses a larger model, so does not run in real time. The output of recognition is a JSON object.
See the face-api.js I site for details.
github.com/justadudewhohacks/face-api.js
Currently implemented: faceapi.detectSingleFace(img) and faceapi.detectSingleFace(img).withFaceLandmarks().
Experiment!