H20 : Human-to-Human-or-Object interaction dataset

H2O dataset is a new interaction dataset dealing with both types of human interactions: Human-to-Human-or-Object. A novel taxonomy of verbs is introduced, to be closer to a description of human body attitude in relation to the surrounding targets of interaction, and more independent of the environment. All interacting instances are annotated with bounding boxes, and distributed in 214 classes. 51 verbs describe the interactions. The annotations of the 14,000 images were made with Pixano.

The dataset has been introduced in this paper: Orcesi, A., Audigier, R., Poka Toukam F. , & Luvison B. (2021, december). Detecting Human-to-Human-or-Object (H2O) Interactions with DIABOLO. In 2021 IEEE International Conference on Automatic Face & Gesture Recognition (FG). IEEE.