Using Clarifai to recognise family members

I currently use Clarifai Face Detection to highlight people in photos, but I cannot see a way to teach Clarifai the faces of family members.

I would like to not only detect the faces, but also guess at the names of the people, hopefully getting better each time I correct it

I see 2 potential ways to address the challenge:

1. Use a workflow: Face detection -> Cropper. This way, you would be generating cropped images from detected faces, and label those images as your training dataset for your classifier. You can then patch the workflow to include the newly trained classifier: Face detection -> Cropper -> Classifier. The idea is laid out in this video: https://www.youtube.com/watch?v=8drVLIaI_iA This would likely be less effort as you're essentially using the detector to crop the images for you. All you need is to label the cropped images.

2. Train a visual-detector from scratch. You would need to draw bounding boxes around the faces AND label them accordingly (by family member).

If you have further questions, let support know via the Clarifai community slack: https://join.slack.com/t/clarifaicommunity/shared_invite/zt-1jehqesme-l60djcd3c_4a1eCV~uPUjQ