Currently, when images are received with a request to train, a directory is created (if it doesn't exist) for a person with the name sent, images are stored in that directory for a period of time, and then training is done inside the MS face container. We should first check to see if a person by that name already exists using the face API and not retrain in that case.
Currently, when images are received with a request to train, a directory is created (if it doesn't exist) for a person with the name sent, images are stored in that directory for a period of time, and then training is done inside the MS face container. We should first check to see if a person by that name already exists using the face API and not retrain in that case.