WIDER FACE Face Detection Benchmark Dataset

WIDER Face Dataset is a face detection benchmark dataset, mainly used for identity verification. It contains 32203 images and 393,703 face images with different scales, poses, standards, expressions, dressing and lighting.

The dataset is organized based on 61 event classes, and for each event class, data from 40%, 10%, and 50% are randomly selected for training, validation, and testing.

The WIDER Face Dataset was released by Yang Shuo, Luo Ping, Loy Chen and Tang Xiaoou from the Chinese University of Hong Kong in 2016. The related paper is "WIDER FACE: A Face Detection Benchmark".