CrowdFix

CrowdFix 包含 434 个具有不同人群场景的视频，总共包含 37,493 帧和 1,249 秒。多样化的内容是指在三个不同类别下的不同人群活动——稀疏、密集自由流动和密集拥挤。所有视频均为 720p 分辨率和 30 Hz 帧速率。为了监测眼球运动，我们的实验中使用了 EyeTribe 眼球追踪器。 26名参与者（10名男性和16名女性），年龄从17岁到40岁，参加了眼动追踪实验。所有参与者都不是眼动追踪实验的专家，视力正常/矫正至正常。在实验过程中，受试者与监视器之间的距离固定为 60 cm。在观看视频之前，每个受试者都需要对眼动仪进行 9 点校准。校准后，受试者被要求免费观看以 MTV 风格显示的视频。最后，为我们的眼动追踪数据库收集了所有 32 名受试者对 538 个视频的注视。

CrowdFix consists of 434 videos depicting various crowd scenarios, with a total of 37,493 frames and a duration of 1,249 seconds. The diverse content covers three distinct categories of crowd activities: sparse, dense free-flowing, and densely crowded scenes. All videos are in 720p resolution with a frame rate of 30 Hz. To monitor eye movements, the EyeTribe eye tracker was used in our experiments. Twenty-six participants (10 males and 16 females) aged from 17 to 40 years old participated in the eye-tracking experiment. All participants were non-experts in eye-tracking experiments, with normal or corrected-to-normal vision. During the experiment, the distance between each participant and the monitor was fixed at 60 cm. Prior to viewing the videos, each participant was required to perform a 9-point calibration for the eye tracker. After calibration, participants were asked to freely watch the videos presented in MTV-style. Finally, gaze data from all 32 participants viewing 538 videos were collected for our eye-tracking database.

搜集汇总

背景与挑战

背景概述

CrowdFix是一个专注于人群场景的眼动追踪数据集，包含434个不同人群活动的视频，涵盖稀疏、密集自由流动和密集拥挤三种场景，所有视频均为720p@30Hz规格。该数据集通过26名非专家参与者的眼动追踪实验收集，用于研究真实人群场景中的视觉注意力模式。