28,972张驾驶员人脸检测&人脸96关键点标注数据

100人驾驶员人脸检测&人脸96关键点标注数据。数据涵盖多年龄段、多时间段、多人种（白种人，黑种人和印度人）驾驶行为包括危险驾驶行为、疲劳驾驶行为和视线偏移行为。在采集设备方面，采用了红外摄像头。在标注方面，一个人包含274-299张照片，标注被采集人的人脸检测框和人脸96关键点。数据可用于驾驶员人脸检测、人脸96关键点识别等任务。

A 100-subject driver face detection and 96 facial landmark annotation dataset. This dataset covers diverse age groups, time periods, and ethnicities (including Caucasian, Black, and Indian populations), and includes various driving behaviors: dangerous driving, fatigued driving, and gaze deviation behaviors. Infrared cameras were utilized for data collection. For annotation, each subject corresponds to 274 to 299 images, with annotated face bounding boxes and 96 facial landmarks of the captured individuals. This dataset can be applied to tasks such as driver face detection and 96 facial landmark recognition.

搜集汇总

背景与挑战

背景概述

该数据集包含28,972张驾驶员人脸图像，标注了人脸检测框和96个面部关键点，数据来自100名不同年龄、种族和驾驶行为（如危险驾驶、疲劳驾驶）的个体。数据通过红外摄像头在车内多种场景下收集，图像格式为.jpg，标注文件为.json，适用于人脸检测和关键点识别任务。