FDDB: A benchmark for face detection in unconstrained settings

By Vidit Jain, Eric Learned-Miller. 2010

1,047 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/cfd1188d9552d54a4bfbd10b611b39bdc6dc7e83

How researchers see this paper

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

Citations in the corpus

A 3D Morphable Model Learnt from 10,000 Faces (2016, background): “[19, 26, 20, 18], and we believe that our publicly available models contributes towards this effort.”
Unsupervised Hard Example Mining from Videos for Improved Object Detection (2018, methodology): “Many of these object detectors have been re-purposed for other tasks such as face detection [39,29,60,15], [31,63,62,25,57,23,66] and pedestrian detection [64,14,6,7,22,30,65], achieving impressive results [24,61,12].”
Recent Advances in Deep Learning for Object Detection (2019, methodology): “This method achieved state-of-the-art results on many public benchmarks such as FDDB [199] and WIDER”
Perceptual Annotation: Measuring Human Vision to Improve Computer Vision (2014, methodology): “FDDB consists of 2,845 images that contain 5,171 annotated faces, split across 10 different folds for cross-validation style testing.”
DenseBox: Unifying Landmark Localization with End to End Object Detection (2015, methodology): “Unlike the widely used FDDB[17] face detection benchmark, which is collected from news photos and the pose tends to be frontal, the face images in MALF have much larger diversity, making it closer to real world application than FDDB. Training and Testing.”
A Survey of Recent Advances in Face Detection (2010, background): “On the other hand, as was pointed in a recent technical report by Jain and Learned-Miller [35], face detection in completely unconstrained settings remains a very challenging task, particularly due to the significant pose and lighting variations.”
R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object (2019, unknown): “g the trade-off of speed and precision, the scale and ratio of the anchor in this paper are set to f20;21=3;22=3gand f1;1=2;2;1=3;3;5;1=5g, respectively. Figure 13: Face detection results on the FDDB [15] benchmarks. Base Model Backbone Anchor Scales Anchor Ratios mAP RetinaNet-H ResNet50 f20g f1;1=”
SSH: Single Stage Headless Face Detector (2017, background): “SSH also achieves state-of-the-art results on the FDDB and Pascal-Faces datasets with a relatively small input size, leading to a runtime of 50 ms/image.”
Max-Margin Object Detection (2015, methodology): “To test the effectiveness of MMOD, we evaluate it on the TU Darmstadt cows [14], INRIA pedestrians [3], and FDDB [8] datasets.”
Detecting Masked Faces in the Wild with LLE-CNNs (2017, background): “In particular, compared with early datasets like FDDB [11], recent datasets are becoming much larger and contain much more real-world scenarios.”
Joint Cascade Face Detection and Alignment (2014, methodology): “For evaluation we use three challenging public datasets: FDDB [9], AFW [32] and CMU-MIT [22].”
Deep Alignment Network: A Convolutional Neural Network for Robust Face Alignment (2017, methodology): “The training subset consists of images from the FDDB [12] and AFLW [15] datasets.”
Convolutional Channel Features (2015, methodology): “We evaluate the CCF face detector on two popular face detection test sets, AFW [54] and FDDB [29].”
Image and Vision Computing (2016, unknown): “Consequently, in order to facilitate the participants and make the competition less dependent to a face detector’s performance, we suggested them to use one of the face detection methods that took part in the Face Detection Data Set and Benchmark (FDDB) [50].”
3D Face Morphable Models "In-the-Wild" (2017, background): “During the past few years, we have witnessed significant improvements in various face analysis tasks such as face detection [19, 42] and 2D facial landmark localisation on static images [40, 21, 6, 38, 43, 4, 5, 36].”
Learn to Combine Multiple Hypotheses for Accurate Face Alignment (2013, methodology): “In face detection field, performance is usually measured by the 50% overlap metric [13], which is not accurate enough for the initialization of face alignment algorithm.”