WIDER FACE: A Face Detection Benchmark

By Shuo Yang, Ping Luo, Chen Change Loy, Xiaoou Tang. 2015

1,809 citations recorded by the source corpus.

Face detection is one of the most studied topics in the computer vision community. Much of the progresses have been made by the availability of face detection benchmark datasets. We show that there is a gap between current face detection performance and the real world requirements. To facilitate future face detection research, we introduce the WIDER FACE dataset1, which is 10 times larger than existing datasets. The dataset contains rich annotations, including occlusions, poses, event categories, and face bounding boxes. Faces in the proposed dataset are extremely challenging due to large variations in scale, pose and occlusion, as shown in Fig. 1. Furthermore, we show that WIDER FACE dataset is an effective training source for face detection. We benchmark several representative detection systems, providing an overview of state-of-the-art performance and propose a solution to deal with large scale variation. Finally, we discuss common failure cases that worth to be further investigated.

Subfields: Computer Vision and Pattern Recognition, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age (2019, background): “To date, numerous large scale face image datasets [22, 31, 13, 69, 37, 24, 41, 68, 15, 27, 46, 8, 39] have been proposed and fostered research and development for automated face detection [35, 21], alignment [66, 44], recognition [57, 49], generation [67, 5, 26, 58], modiﬁcation [3, 32, 19], and…”
Cascade R-CNN: High Quality Object Detection and Instance Segmentation (2019, methodology): “Finally,we further present the results of a larger set of experiments, performed on various popular generic/specific object detection datasets, including PASCAL VOC [12], KITTI [15], CityPerson [63] and WiderFace [60].”
Deep High-Resolution Representation Learning for Visual Recognition (2019, methodology): “The WFLW dataset [149] is a recently-built dataset based on the WIDER Face [164].”
Condensed Movies: Story Based Retrieval with Contextual Embeddings (2020, methodology): “Face embeddings for each face track are extracted in three stages: (1) Each frame is passed through a dual shot face detector [54] (trained on the Wider Face dataset [55]) to extract bounding boxes.”
IARPA Janus Benchmark-B Face Dataset (2017, background): “/association_modified_bcubed.py detection_cluster_file ground_truth [-no_fte][-h] – detection_cluster_file is the output of clustering Test 8 and should have a header of: TEMPLATE_ID,FILENAME,CLUSTER_INDEX,CONFIDENCE,FACE_X,FACE_Y,FACE_WIDTH,FACE_HEIGHT – ground_truth associates the detected boundin”
Detecting small faces in the wild based on generative adversarial network and contextual information (2019, methodology): “In this section, we experimentally validate our proposed method on two public face detection benchmarks ( i.e. WIDER FACE [25] and FDDB [24] ).”
Objects365: A Large-Scale, High-Quality Dataset for Object Detection (2019, background): “Besides the general object detection datasets, there are also a lot of other detection benchmarks like face detection [19, 35], pedestrian detection [7, 36, 31], and human/vehicle detection for the autonomous driving [11, 2], all of which play an important role in the detection community.”
Look at Boundary: A Boundary-Aware Face Alignment Algorithm (2018, methodology): “WFLW dataset: In order to facilitate future research of face alignment, we introduce a new facial dataset base on WIDER Face [61] named Wider Facial Landmarks in-the-wild (WFLW), which contains 10000 faces (7500 for training and 2500 for testing) with 98 fully manual annotated landmarks.”
SSDMNV2: A real time DNN-based face mask detection system using single shot multibox detector and MobileNetV2 (2020, positive): “Earlier datasets consisted of images fetched in supervised surroundings, while recent datasets are constructed by taking online images like WiderFace (Yang, Luo, Loy, & Tang, 2016), IJB-A (Klare et al., 2015a), MALF (Yang, Yan, Lei, & Li, 2015), and CelebA (Klare et al., 2015b).”
UA-DETRAC: A new benchmark and protocol for multi-object detection and tracking (2015, methodology): “For evaluation on object detection, similar to the KITTI detection (Geiger et al., 2012) and WIDER FACE (Yang et al., 2016) datasets, we deﬁne three levels of di ﬃ culties in the UA-DETRAC-test set, i.e. , easy (10 sequences), medium (20 sequences), and hard (10 sequences) based on the recall rate…”
From Facial Expression Recognition to Interpersonal Relation Prediction (2016, background): “tion [84,85,42,82,55], face parsing [50,47], face landmark detection [92,99,73], face attribute recognition [48,78,26], face recognition [66,58,68], and face clustering [94].”
Benchmarking Low-Light Image Enhancement and Beyond (2021, methodology): “The baseline methods are trained on WIDER FACE (Yang et al. 2016), a large dataset captured in the well-exposed condition with large scale variations in diversiﬁed attributes and conditions.”
IARPA Janus Benchmark - C: Face Dataset and Protocol (2018, background): “WIDER FACE, a large scale face detection dataset released in 2016, made significant strides towards addressing the data quantity problem associated with evaluation of face detection algorithms [19].”
UnitBox: An Advanced Object Detection Network (2016, methodology): “The weights of UnitBox are initialized from a VGG-16 model pre-trained on ImageNet, and then ﬁne-tuned on the public face dataset WiderFace [14].”
Unsupervised Domain Adaptation of Object Detectors: A Survey (2021, background): “These datasets can be classiﬁed into the following categories: (1) General objects [19], [99], (2) Self-driving [172], [173], [174], [175], (3) Face detection [176], [177], (4) Weather degradation [174], [175], [176], [178], and (5) Synthetic datasets [179].”
SSH: Single Stage Headless Face Detector (2017, methodology): “For face detection, CMS-RCNN [38] used the same idea as [1, 36] and added skip connections to the Faster RCNN [24].”
Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement (2020, methodology): “A state-of-the-art deep face detector, Dual Shot Face Detector (DSFD) [14], trained on WIDER FACE dataset [28], is used as the baseline model.”