Dataset Groups Activity Stream AVA v2.2 The AVA v2.2 dataset for spatiotemporal action localization contains the bounding box annotations and the corresponding action labels on keyframes. BibTex: @dataset{Zhifan_Wan_and_Jie_Zhang_and_Changzhen_Li_and_Shiguang_Shan_2024, abstract = {The AVA v2.2 dataset for spatiotemporal action localization contains the bounding box annotations and the corresponding action labels on keyframes.}, author = {Zhifan Wan and Jie Zhang and Changzhen Li and Shiguang Shan}, doi = {10.57702/6p68znnc}, institution = {No Organization}, keyword = {'action recognition', 'atomic visual actions', 'ava-v2-2', 'spatio-temporal localization', 'video action localization', 'video analysis'}, month = {dec}, publisher = {TIB}, title = {AVA v2.2}, url = {https://service.tib.eu/ldmservice/dataset/ava-v2-2}, year = {2024} }
