数据堂—180人9,000张驾驶员手势21关键点标注数据

180人9,000张驾驶员手势21关键点标注数据。数据涵盖多年龄段、多种时间段、多种手势、多种车型、多种时间段。在标注方面，标签标注车型、手势类别、被采集人国籍、性别和年龄；标注驾驶员手势21关键点（每个关键点有可见不可见属性）。数据可用于驾驶员手势识别、手势关键点检测和识别等任务。

This is a 21-keypoint annotated driver gesture dataset containing 9,000 images from 180 individuals. The dataset covers diverse age groups, various time periods, different gesture categories, multiple vehicle models, and multiple time periods. For annotation, the dataset provides labels for vehicle model, gesture category, nationality, gender and age of the subjects, as well as 21 keypoints of driver gestures where each keypoint has a visibility attribute. This dataset can be used for tasks including driver gesture recognition, gesture keypoint detection and recognition.

搜集汇总

背景与挑战

背景概述

该数据集包含来自180人的9,000张驾驶员手势图像，每张图像标注了21个关键点，并覆盖了多种年龄、时间段、手势和车型。标注信息包括车型、手势类别、人员国籍、性别和年龄，可用于驾驶员手势识别和关键点检测任务。数据采集自车内摄像头，图像格式为.jpg，标注文件为.json，关键点标注准确率不低于95%。