The reason why use opencv for preprocess_face

[KIMS0YOUNG]

Hi, thanks you for amazing work. By the way, I have a question for realtime.py's analysis function.

for face recognition, custom_face has defined by functions.preprocess_face() in realtime.py line 340. but in that code, you used fixed detector_backend='opencv' not detector_backend=detector_backend. So when I use DeepFace.stream("database", detector_backend='retinaface'), custom_face's detector_backend isn't chaged. Is there any reason why you used fixed detector_backend for preprocess_face() ?

Thanks.

[serengil]

hey, i think you are using out-of-the-date code

https://github.com/serengil/deepface/blob/master/deepface/commons/realtime.py

this one is not pubished on pip yet, you can use the source code instead of pypi version

[KIMS0YOUNG]

Thanks a lot.