I am trying to do face recognition on an ongoing video stream. Detected faces will be checked if they match any already trained label, if so the model will be updated, if not a new label will be started. My fear now is, as people come and go, more and more labels will be created and memory is wasted on people who have left the scene and never come back. Is there a way to "untrain" a certain label and release the memory?
This forum is disabled, please visit https://forum.opencv.org