EVIP：基于眼部视频的非接触式眼压测量数据集

数据集获取（Figshare）

• 源站标题：EVIP: An eye video-based intraocular pressure dataset for non-contact intraocular pressure measurement
• DOI：10.6084/m9.figshare.29605745.v3
• 发布日期：2025-07-25
• 文件数：183 个
• 文件总体积：2.46 GB
• 源站页面：https://figshare.com/articles/dataset/EVIP_An_eye_video-based_intraocular_pressure_dataset_for_non-contact_intraocular_pressure_measurement/29605745

命令行下载

curl -L -o "1.MP4" "https://ndownloader.figshare.com/files/56406074"
curl -L -o "10.mp4" "https://ndownloader.figshare.com/files/56405687"
curl -L -o "100.mp4" "https://ndownloader.figshare.com/files/56405567"
curl -L -o "101.mp4" "https://ndownloader.figshare.com/files/56405573"
curl -L -o "102.mp4" "https://ndownloader.figshare.com/files/56405738"
curl -L -o "103.mp4" "https://ndownloader.figshare.com/files/56405570"
curl -L -o "104.mp4" "https://ndownloader.figshare.com/files/56405582"
curl -L -o "105.mp4" "https://ndownloader.figshare.com/files/56405588"

文件清单

• 1.MP4（87.55 MB）
• 10.mp4（8.27 MB）
• 100.mp4（1.93 MB）
• 101.mp4（3.80 MB）
• 102.mp4（8.22 MB）
• 103.mp4（1.41 MB）
• 104.mp4（2.62 MB）
• 105.mp4（3.57 MB）
• 106.mp4（2.16 MB）
• 107.mp4（2.19 MB）
• 108.mp4（1.68 MB）
• 109.mp4（2.07 MB）
• 11.mp4（5.58 MB）
• 110.mp4（2.18 MB）
• 111.mp4（8.82 MB）
• 112.mp4（8.66 MB）
• 113.mp4（4.15 MB）
• 114.mp4（3.19 MB）
• 115.mp4（4.32 MB）
• 116.mp4（3.49 MB）

文件总体积 2.46 GB（源站 API 实测）。

数据加载示例（图像类）

from PIL import Image
import glob, os

files = (glob.glob(os.path.join(path, "**", "*.png"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.jpg"), recursive=True)
       + glob.glob(os.path.join(path, "**", "*.tif"), recursive=True))
print("图像文件数:", len(files))
img = Image.open(files[0]); print("尺寸/模式:", img.size, img.mode)

# torchvision Dataset 方式:
# from torchvision import datasets
# ds = datasets.ImageFolder(path)  # 要求 子目录=类别

目录组织与标注格式以源站说明和下载后实际文件为准。

数据说明

Existing facial rPPG datasets (e.g., for heart rate or oxygen saturation monitoring) primarily capture full-face regions, focusing on skin perfusion signals from the forehead, cheeks, or periorbital areas. These datasets lack specialized annotations for ocular structures (pupil, iris, sclera) and direct IOP measurements, as their core goal is general physiological parameter estimation rather than eye-specific pressure analysis. In contrast, EVIP is uniquely designed to link ocular physiological signals with IOP values: it provides high-resolution eye-region videos (isolating pupil-iris complexes) paired with clinically validated IOP measurements, enabling the extraction of IOP-related BVP features from eye regions.
The folder contains two subfolders, EVIP-1 and EVIP-2, representing the two versions of the EVIP dataset. Each subfolder contains eye videos recorded in MP4 format, along with data files containing the actual IOP values recorded and stored in CSV file. For the EVIP-1 folder, there were 60 eye videos from 30 participants, named in Arabic numeral order, along with an “EVIP-1.CSV” data file. The data included a column named “Actual Value” containing the actual IOP records, which corresponded to the order of the video files from top to bottom. EVIP-2 was stored in the same manner, containing 122 eye videos from 60 participants, along with a synchronously recorded actual IOP value data file named “EVIP-2.CSV”.

许可

CC BY 4.0

引用

学术使用请引用 DOI 10.6084/m9.figshare.29605745.v3（source: https://figshare.com/articles/dataset/EVIP_An_eye_video-based_intraocular_pressure_dataset_for_non-contact_intraocular_pressure_measurement/29605745）。