Face-in-Action 面部视频数据集

Face-in-Action Dataset 是具有成像变化的面部视频数据集，其涵盖来自 180 名参与者的 20 秒面部视频数据，主要用于模仿护照检查方案，相关视频由 3 个不同角度的 6 个同步摄像机捕获（每个角度的焦距为 8mm 和 4mm），图像采集于受限的室内环境和开放的室外环境。

The Face-in-Action Dataset is a facial video dataset with imaging variations. It contains 20-second facial video data from 180 participants, and is primarily designed for simulating passport inspection scenarios. The related videos are captured by 6 synchronized cameras across 3 distinct angles, with each angle equipped with cameras featuring focal lengths of 8mm and 4mm. Image acquisition is conducted in both constrained indoor environments and unconstrained outdoor environments.

搜集汇总

背景与挑战

背景概述

Face-in-Action 是一个面部视频数据集，包含180名参与者在室内外环境下拍摄的20秒视频，用于模拟护照检查场景，视频由6个同步摄像机从不同角度捕获。该数据集由卡耐基·梅隆大学于2010年发布，遵循非商业许可证，适用于安防和机器视觉领域。