Segment objects from images using natural language prompts
Visualize egocentric and exocentric human activity datasets