IReye dataset, HMD Face
收藏资源简介:
IReye数据集是通过Fove 0 HMD收集的,包含7个主题。使用Fove的调试工具和屏幕录像机收集数据。每次数据收集前,都会使用默认程序校准用户眼睛,然后收集图像和标签。校准需要10-20秒,屏幕右下角会显示校准已结束。数据收集有两种方式:1. 使用Fove提供的调试工具记录图像和标签,然后使用OCR技术提取数字。2. 使用当前的Fove SDK API直接收集图像和标签。
The IReye dataset was collected using the Fove 0 HMD and includes data from seven subjects. Data collection was facilitated by Fove's debugging tools and screen recording software. Prior to each data collection session, the user's eyes were calibrated using a default procedure, followed by the collection of images and labels. The calibration process took between 10 to 20 seconds, with a notification displayed in the lower right corner of the screen indicating its completion. Data was collected in two ways: 1. Using Fove's provided debugging tools to record images and labels, followed by the extraction of numbers using OCR technology. 2. Directly collecting images and labels using the current Fove SDK API.
EyeFaceDataset
- 来源: 数据集链接为https://drive.google.com/drive/folders/1PWq3gJKoDo6m-ANgl9erqEbUuYn0Ihts,用于论文"Real-time 3D Face-Eye Performance Capture of a Person Wearing VR Headset"。
IReye dataset
- 采集设备: Fove 0 HMD。
- 参与者数量: 7名。
- 数据收集方法:
- 使用Fove的调试工具记录图像和标签,结合OCR技术提取数字。
- 使用Fove SDK API直接收集图像和标签。
- 校准过程: 需要10-20秒,屏幕右下角显示"Calibration has ended."。
- 瞳孔位置追踪: 参考https://github.com/YutaItoh/3D-Eye-Tracker,需调整常规尺寸和参数。
- 瞳孔大小测量: 使用经典的分水岭分割方法提取瞳孔区域并确定其大小。
HMD Face
- 采集设备: Mac摄像头。
- 3DMM参数提取: 使用CoarseNet,代码来源https://github.com/Juyong/3DFace。
- 3D面网格创建: 使用"Align.py"代码。
- HMD与面部对齐: 提供"HMDgh.gh"文件,为Rhino5的Grasshopper脚本,需下载Rhino5和Grasshopper插件,并调整HMD与面部的距离和HMD的缩放比例。
联系方式
- 邮箱: guoxian001@e.ntu.edu.sg




