Sensing, Recognition, and Security / Eye Gaze Tracking Sensing

In crowded multi-object scenes, how can activity knowledge graphs represent "what a person is looking at" relationships?

Similar questions

Related papers