No. Our model is pre-trained with tens of millions of images, it's plug and play via our APIs/SDKs. You only need to register a person with single image and it is also sufficient to recognize the person with a single image.