Specifies what you want to detect in the video, such as people, packages, or pets. The current valid labels you can include in this list are: "PERSON", "PET", "PACKAGE", and "ALL".
The minimum confidence required to label an object in the video.