Use CoreML on detected faces

[rsmithde]

I am trying to use CoreML on the detected faces. My original idea was to capture the image of each detected face and then run that agains the models, but I can't figure out how to get the screenshot of each face. Any hints? I tried using the facebounds in the drawFaceboundingBox method and using that rect to crop the image, but they don't line up.

[Willjay90]

IMO, using the bounding box returned by the API and you can get the screenshot by this method according to your desired size.

[jainamSynoverge]

IMO, using the bounding box returned by the API and you can get the screenshot by this method according to your desired size.

is there any way through we can compare faces after face detection like face verification?

[Willjay90]

@jainamSynoverge you have to train another network to do the verification

[jainamSynoverge]

@jainamSynoverge you have to train another network to do the verification

@Willjay90 what do you mean by another network, can you plz explain in some detail?