I tried inference model with webcam BGR input frame, use SCRFD to detect face then crop and convert to GRAY image before feed to the FAU model but the prediction is not consistent with my movement. Could you please share the correct way to preprocess image?
I tried inference model with webcam BGR input frame, use SCRFD to detect face then crop and convert to GRAY image before feed to the FAU model but the prediction is not consistent with my movement. Could you please share the correct way to preprocess image?