Segment objects in images using text prompts
Visualize egocentric and exocentric human activity datasets