Labeled Faces in the Wild : Updates and New Reporting Procedures

By Gary B. Huang, Eric Learned-Miller. 2014

433 citations recorded by the source corpus.

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

Machine Learning Models that Remember Too Much (2017, background): “Labeled Faces in the Wild (LFW) contains 13,233 images for 5,749 individuals [33, 45].”
I-Pic: A Platform for Privacy-Compliant Image Capture (2016, methodology): “Per-user SVM classifiers are then trained on the FNet features, where positive examples consist of the portraits provided by the corresponding user, and negative examples from the other users and ∼12K celebrity faces in the Labeled Faces in the Wild dataset (LFW) [17].”
SphereFace: Deep Hypersphere Embedding for Face Recognition (2017, methodology): “We follow the unrestricted with labeled outside data protocol [8] on both datasets.”
MS-Celeb-1M: A Dataset and Benchmark for Large-Scale Face Recognition (2016, methodology): “Available people images IJB-A [17] public 500 5712 LFW [7,8] public 5K 13K YFD [14] public 1595 3425 videos CelebFaces [15] public 10K 202K CASIA-WebFace [16] As shown in Table 1, our training dataset is considerably larger than the publicly available datasets.”
Web-scale training for face identification (2014, methodology): “As the verification protocol, we follow the LFW unrestricted protocol [9] (which uses only the same-not-same labels), and similarly to [22] train a kernel SVM (with C=1) on top of the χ(2)-distance vectors derived from the computed representations.”
One-shot Face Recognition by Promoting Underrepresented Classes (2017, methodology): “The base set provided with this paper is mainly to train a robust, generalizable face feature extractor and isolate the problem of one-shot learning from feature Dataset Available people images IJB-A [11] public 500 5712 LFW [9, 8] public 5 K 13 K YFD [22] public 1595 3425 videos CelebFaces [17] pub”
Learning Compact Binary Face Descriptor for Face Recognition (2015, background): “There are six evaluation paradigms on this dataset [20].”
Slim-CNN: A Light-Weight CNN for Face Attribute Prediction (2019, methodology): “In this work, we use two publicly available datasets: CelebA[19] and LFWA dataset[15].”
A model explanation system (2016, methodology): “Face recognition example We now demonstrate the black box MES implementation on the sci-kit learn demo “Faces recognition example using eigenfaces and SVMs” [7; 17].”
Labeled Faces in the Wild: A Survey (2016, methodology): “These protocols were described in a second technical report [47].”
Pruning and Quantization for Deep Neural Network Acceleration: A Survey (2021, background): “52% accuracy loss on labeled faces in the wild (LFW) dataset [104] in the filed of face recognition.”
Unsupervised Object Segmentation by Redrawing (2019, neutral): “Labeled Faces in the Wild dataset [25, 33] is a dataset of 13233 faces.”
Face/Off: Preventing Privacy Leakage From Photos in Social Networks (2015, background): “com and according to a comparative study [28], the face recognition algorithm of face.”
NormFace: L2 Hypersphere Embedding for Face Verification (2017, methodology): “We evaluate our methods through two di erent protocols on LFW, one is the standard unrestricted with labeled outside data [9], which is evaluated on 6, 000 image pairs, and another is BLUFR [15] which utilize all 13, 233 images.”
Targeting Ultimate Accuracy: Face Recognition via Deep Embedding (2015, methodology): “Please refer to [5, 3, 13, 14 ] for the details of all the protocols.”
A Comprehensive Study on Center Loss for Deep Face Recognition (2019, methodology): “For both LFW and YTF benchmarks, we follow the unrestricted with labeled outside data protocol (Huang and Learned-Miller 2014).”