Heightened Reality Games
Scenario 1: Your friend may be an evil impersonator.
Input Description:
Smiling woman with short dark hair, wearing glasses, a purple butterfly t-shirt, purple necklace, and spiral earrings
Derived Keywords:
[Smiling woman with short dark hair]
[wearing glasses]
[a purple butterfly t-shirt]
[purple necklace]
[and spiral earrings]
Using YOLO-World Machine Vision, we have onboard image and video classification.
YOLO-World uses natural language processing, enabling a rapid search of the derived keywords from the player.
Because of its relatively small size and hyper-fast zero-shot detection, this is onboard and requires no cloud/server interaction.
Despite lower accuracy, we can prime the player to provide useful keywords that are easy to detect.