数据堂—2,341人会议场景人脸多姿态数据

2,341人会议场景人脸多姿态数据，包含黄种人、白种人、黑种人、棕色人种，年龄以中青年为主，采集了多种室内办公场景，涵盖会议室、咖啡厅、图书馆、卧室等，每人采集了23段视频和4张图像，视频包含张嘴，转头、闭眼、手摸耳朵等23种姿态，图像包含戴口罩、戴墨镜等四种姿态

A multi-pose face dataset with 2,341 individuals captured in meeting scenarios. The dataset includes four major racial groups: Mongoloid (yellow-skinned), Caucasian (white-skinned), Negroid (black-skinned) and brown-skinned populations, with most subjects being young and middle-aged adults. Data collection was conducted across various indoor office-related environments, including conference rooms, cafes, libraries, bedrooms and other similar spaces. For each participant, 23 video segments and 4 still images were collected. The videos cover 23 types of poses and behaviors such as opening the mouth, turning the head, closing the eyes, touching ears with hands, etc. The still images feature four different poses including wearing face masks and sunglasses.

搜集汇总

背景与挑战

背景概述

该数据集包含2,341人的会议场景人脸多姿态数据，每人有23个视频片段和4张图像，涵盖亚洲、高加索、非洲和棕色人种，主要聚焦中青年群体。数据采集于室内办公环境，如会议室和咖啡馆，并包含多种姿态变化，如张嘴、转头和戴口罩等。