AGD20K

The AGD20K dataset provides a collection of 20,061 images captured from a third-person perspective (exocentric) and 3,755 images from a first-person perspective (egocentric). These images are annotated with labels for 36 different affordances, which represent the potential object interactions.

Data and Resources

Cite this as

Claudia Cuttano, Gabriele Rosi, Gabriele Trivigno, Giuseppe Averta (2024). Dataset: AGD20K. https://doi.org/10.57702/dyj1ohrk

DOI retrieved: December 16, 2024

Additional Info

Field Value
Created December 16, 2024
Last update December 16, 2024
Defined In https://doi.org/10.48550/arXiv.2404.12015
Author Claudia Cuttano
More Authors
Gabriele Rosi
Gabriele Trivigno
Giuseppe Averta
Homepage https://arxiv.org/abs/2106.09528