Gaze Data for the Hollywood2 Action Recognition Data Set

Gaze data from 5 subjects watching the Hollywood2 Action Recognition Dataset while performing the action recognition task set out for the challenge. from: Eleonora Vig, Michael Dorr, David Cox (2012) Space-Variant Descriptor Sampling for Action Recognition Based On Saliency and Eye Movements. European Conference on Computer Vision本数据集收录了5名受试者在观看Hollywood2动作识别数据集（Hollywood2 Action Recognition Dataset）并执行挑战赛预设的动作识别任务过程中采集的眼动数据。数据来源：Eleonora Vig、Michael Dorr、David Cox（2012）发表于欧洲计算机视觉大会（European Conference on Computer Vision）的论文《基于显著性与眼动的动作识别空间变描述符采样》（Space-Variant Descriptor Sampling for Action Recognition Based On Saliency and Eye Movements）。