NIST人脸数据集（IJB）

由于美国国家标准化研究院（NIST）发布的大型人脸数据集，包括从互联网采集的静态人脸图像和视频，共有1845个对象，11754张图片，55026视频帧，7011个视频和10044非人脸图像。该数据集在采集上与其它流行的人脸数据集不重叠，如：牛津大学VGG人脸数据集Face Challenge、CASIA WEBFace数据集等。 数据集被用于NIST举办的 Face Challenge 竞赛，进行校验、搜索、检测、聚类、标注等人脸识别任务。

The large-scale facial dataset released by the National Institute of Standards and Technology (NIST) is composed of static facial images and videos collected from the Internet, with 1845 distinct subjects, 11754 facial images, 55026 video frames, 7011 video clips, and 10044 non-face images in total. This dataset does not overlap with other widely used face datasets in terms of data collection, including the Oxford University VGG Face Challenge dataset and the CASIA WEBFace dataset. It was utilized in the Face Challenge competition hosted by NIST to support various face recognition tasks such as verification, retrieval, detection, clustering and annotation.

搜集汇总

背景与挑战

背景概述

NIST人脸数据集（IJB）是一个由美国国家标准化研究院发布的大型人脸数据集，包含1845个对象、11754张图片、55026视频帧和7011个视频，主要用于人脸识别任务。该数据集与其他流行的人脸数据集不重叠，适用于多种人脸识别应用场景。