Egocentric Vision-based Action Recognition: A survey

The egocentric action recognition EAR field has recently increased its popularity due to the affordable and lightweight wearable cameras available nowadays such as GoPro similars. Therefore, amount of data generated increased, triggering interest in understanding videos. More specifically, actions videos gained challenge that it poses: wild movement camera lack context make hard recognise with a performance similar third-person vision solutions. This ignited research on and, nowadays, many public datasets competitions can be found both machine learning computer communities. In this survey, we aim analyse literature methods algorithms. For that, propose taxonomy divide into various categories subcategories, contributing more fine-grained classification methods. We also provide review zero-shot approaches used by community, methodology could help transfer algorithms real-world applications. Finally, summarise researchers literature.