TEyeD

我们展示了TEyeD，这是世界上最大的使用头戴式设备拍摄的眼睛图像的统一公共数据集。TEyeD是用七个不同的头戴式眼部追踪器获得的。其中，两个眼睛跟踪器被集成到虚拟现实 (VR) 或增强现实 (AR) 设备中。TEyeD中的图像是从各种任务中获得的，包括汽车骑行，模拟器骑行，户外体育活动和日常室内活动。数据集包括所有图像的2D和3D地标，语义分割，3D眼球注释以及凝视矢量和眼睛运动类型。为瞳孔，虹膜和眼睑提供了地标和语义分割。视频长度从几分钟到几小时不等。TEyeD拥有2000万多个经过精心注释的图像，为在现代VR和AR应用中推进计算机视觉，眼睛跟踪和凝视估计领域的研究提供了独特，连贯的资源和宝贵的基础。

We present TEyeD, the largest unified public dataset of eye images captured via head-mounted devices worldwide. TEyeD was collected using seven distinct head-mounted eye trackers, two of which are integrated into virtual reality (VR) or augmented reality (AR) devices. Images in TEyeD are acquired from a wide range of tasks, including car driving, simulator rides, outdoor sports activities, and daily indoor activities. The dataset provides 2D and 3D landmarks, semantic segmentation annotations, 3D eyeball annotations, as well as gaze vectors and eye movement types for all images. Landmarks and semantic segmentation masks are provided for the pupil, iris, and eyelids. The source videos have durations ranging from several minutes to multiple hours. TEyeD contains over 20 million carefully annotated images, serving as a unique, coherent resource and valuable foundation for advancing research in the fields of computer vision, eye tracking, and gaze estimation in modern VR and AR applications.

搜集汇总

背景与挑战

背景概述

TEyeD是一个公开的智慧医疗数据集，托管在OpenDataLab平台，数据集大小为1.1KB，关注量较高（2.2k），但详情页未提供具体数据内容或应用描述。