数据堂—2,341人会议场景手势识别数据

2,341人会议场景手势识别数据，包含黄种人、白种人、黑种人、棕色人种多任务，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18张图片和2段视频，图片包含单手握拳，单手比心等18个手势动作，视频包含鼓掌等手势动作

A gesture recognition dataset for conference scenarios consisting of 2,341 participants. It supports multi-task learning across four racial groups: Yellow, White, Black and Brown. Most participants are young and middle-aged individuals. The dataset was collected in various indoor office environments, including conference rooms, cafes, libraries, bedrooms and other similar spaces. Each participant provided 18 images and 2 video clips. The images include 18 gesture actions such as single-handed fist clenching and single-handed heart gesture, while the video segments contain gesture actions like clapping.

搜集汇总

背景与挑战

背景概述

该数据集包含2,341人的手势识别数据，涵盖多种族、年龄和性别，主要针对会议场景。数据包括每人18张图像和2个视频片段，收集于室内办公环境，支持手势识别测试任务。