Researchers from the University of Bristol and the University of Catania in Italy have released a dataset of 100 hours of video recordings of individuals performing activities in the kitchen. The recordings include more than 20,000 unique narrations for the tasks, such as cutting peppers, an individual is performing. The researchers collected the recordings using cameras on individuals’ heads in 45 different kitchens. The dataset can help test the ability of AI systems to detect and anticipate tasks.
Teaching AI Systems to Analyze First-Person Videos
previous post