7 datasets found

The dataset used in this paper for training and testing the legged locomotion system using egocentric vision.

Ego-4D dataset

Epic Kitchens dataset is a dataset for egocentric vision and action recognition.

EPIC-KITCHENS is a large-scale egocentric video benchmark recorded by 32 participants in their native kitchen environments. Our videos depict non-scripted daily activities: we...

The EGTEA Gaze+ dataset offers approximately 10,000 samples of 106 non-scripted daily activities that occur in a kitchen.

EPIC-KITCHENS-55 is a dataset of 432 videos acquired by 32 subjects and labeled with 39,595 action segments.

Egocentric action anticipation is the task of predicting the future actions a camera wearer will likely perform based on past video observations.