Possible future feature?
#3
by
Gelaneier
- opened
I'm thinking if there is a chance that we can add a "gaze" feature into this model--after gazing at a object for like 3 seconds, it start to pass the current frame into a specialized object recognition model so that it could get a better result.