Dataset Groups Activity Stream V-COCO The V-COCO dataset contains 2,533 training images, 2,867 validation images, and 4,946 test images, including 24 action classes. BibTex: @dataset{Frederic_Z_Zhang_and_Yuhui_Yuan_and_Dylan_Campbell_and_Zhuoyao_Zhong_and_Stephen_Gould_2024, abstract = {The V-COCO dataset contains 2,533 training images, 2,867 validation images, and 4,946 test images, including 24 action classes.}, author = {Frederic Z. Zhang and Yuhui Yuan and Dylan Campbell and Zhuoyao Zhong and Stephen Gould}, doi = {10.57702/dlw81x1i}, institution = {No Organization}, keyword = {'Computer Vision', 'Human-Object Interaction', 'Image Dataset', 'Image Processing', 'Object Detection', 'human-object interaction', 'scene graph', 'visual features'}, month = {dec}, publisher = {TIB}, title = {V-COCO}, url = {https://service.tib.eu/ldmservice/dataset/v-coco}, year = {2024} }
