Papperoni

2014

Deepface: Closing the Gap to Human-Level Performance in Face Verification

Y. Taigman, Michael Yang, Marc'aurelio Ranzato, Lior Wolf

Open PDF Google Scholar

citations

Cite Score

82

AI summary

This paper introduces DeepFace, a face verification system using a nine-layer deep neural network trained on a dataset of four million facial images and explicit 3D face modeling, achieving a 97.35% accuracy on the LFW dataset, approaching human-level performance.

Main Contributions

Developed an effective deep neural net (DNN) architecture and learning method that leverages a very large labeled dataset of faces in order to obtain a face representation that generalizes well to other datasets
Introduced an effective facial alignment system based on explicit 3D modeling of faces
Achieved state of the art results on the Labeled Faces in the Wild benchmark (LFW), reaching near human-performance
Achieved state of the art results on the YouTube Faces dataset (YTF), decreasing the error rate there by more than 50%

Abstract

In modern face recognition, the conventional pipeline consists of four stages: detect ⇒ align ⇒ represent ⇒ classify. We revisit both the alignment step and the representation step by employing explicit 3D face modeling in order to apply a piecewise affine transformation, and derive a face representation from a nine-layer deep neural network. This deep network involves more than 120 million parameters using several locally connected layers without weight sharing, rather than the standard convolutional layers. Thus we trained it on the largest facial dataset to-date, an identity labeled dataset of four million facial images belonging to more than 4,000 identities. The learned representations coupling the accurate model-based alignment with the large facial database generalize remarkably well to faces in unconstrained environments, even with a simple classifier. Our method reaches an accuracy of 97.35% on the Labeled Faces in the Wild (LFW) dataset, reducing the error of the current state of the art by more than 27%, closely approaching human-level performance.

Citation Graph

Loading graph...

References [32]

Sort:

Filter:

[1]ImageNet Classification With Deep Convolutional Neural Networks

Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton - 2012

71 papers in library cite

I'm giving this a 5 just because of the impact, but this is VEEERY derivative of earlier work. Kudos for them for putting it all together, but really there's nothing revolutionary here.

[2]Gradient-Based Learning Applied to Document Recognition

Yann Lecun, Leon Bottou, Yoshua Bengio, Patrick Haffner - 1998

62 papers in library cite

I absolutely hated this paper. Has ~50 pages but seems like 200 pages. Takes too long to explain some things that really is just repeating itself. Also doesn't seem to add too much on top of LeNet-5. Also, focuses a lot on GTNs, which really didn't stick.

[3]Learning Representations by Back-Propagating Errors

D. E. Rumelhart, Geoffrey E. Hinton, Ronald J. Williams - 1986

34 papers in library cite

Introduced backprop. Short and simple.

[4]Learning Deep Architectures for AI

Yoshua Bengio - 2009

25 papers in library cite

It's a nice overview. Some sections get very theoretical, but the first half is very good and I feel that it does a waaaay better job of explaining RBMs and DBNs than other papers. This feels like Bengio is taking your hand and saying "if you don't know what's going on, here you go, everything you need to know to jump into the deep nets train"

[5]Large Scale Distributed Deep Networks

Jeffrey Dean, G. S. Corrado, R. Monga, K. Chen, M. Devin, Quoc V. Le, Mark Z. Mao, Marc'aurelio Ranzato, A. Senior, P. Tucker, K. Yang, Andrew Y. Ng - 2012

16 papers in library cite

Good paper, nice algorithm. Nothing too crazy, but I understand the impact. I think the work to create the system was larger than the algorithm itself.

[6]Labeled Faces in the Wild: A Database for Studying Face Recognition in Unconstrained Environments

G. B. Huang, M. Ramesh, T. Berg, E. L. Miller - 2007

5 papers in library cite

Seems relevant but I am tired of reading papers about datasets, especially the early ones.

[7]Synergistic Face Detection and Pose Estimation With Energy-Based Models

M. Osadchy, Yann Lecun, M. Miller - 2007

4 papers in library cite

[8]Emergence of Complex-Like Cells in a Temporal Product Network With Local Receptive Fields

K. Gregor, Yann Lecun - 2010

3 papers in library cite

[9]Attribute and Simile Classifiers for Face Verification

N. Kumar, A. C. Berg, P. N. Belhumeur, S. K. Nayar - 2009

2 papers in library cite

[10]Improving Deep Neural Networks for LVCSR Using Rectified Linear Units and Dropout

George E. Dahl, T. N. Sainath, Geoffrey E. Hinton - 2013

2 papers in library cite

[11]Learning a Similarity Metric Discriminatively, With Application to Face Verification

S. Chopra, Raia Hadsell, Yann Lecun - 2005

2 papers in library cite

[12]A Practical Transfer Learning algorithm for Face Verification

X. Cao, D. Wipf, F. Wen, G. Duan, Jian Sun - 2013

1 paper in library cites

Missing author list

[13]An Introduction to the Good, the Bad, & the Ugly Face Recognition Challenge Problem

2011

1 paper in library cites

[14]Bayesian Face Revisited: A Joint Formulation

Deli Chen, X. Cao, Lisa Wang, F. Wen, Jian Sun - 2012

1 paper in library cites

[15]Blessing of Dimensionality: High-Dimensional Feature and Its Efficient Compression for Face Verification

Deli Chen, X. Cao, F. Wen, Jian Sun - 2013

1 paper in library cites

[16]Deep Convolutional Network Cascade for Facial Point Detection

Y. S. Sun, Xinpeng Wang, X. Tang - 2013

1 paper in library cites

[17]Face Description With Local Binary Patterns: Application to Face Recognition

T. Ahonen, A. Hadid, M. Pietikainen - 2006

1 paper in library cites

[18]Face Recognition in Unconstrained Videos With Matched Background Similarity

Lior Wolf, T. Hassner, I. Maoz - 2011

1 paper in library cites

[19]Fast High Dimensional Vector Multiplication Face Recognition

O. Barkan, J. Weill, Lior Wolf, H. Aronowitz - 2013

1 paper in library cites

[20]Fisher Vector Faces in the Wild

K. Simonyan, O. M. Parkhi, A. Vedaldi, Andrew Zisserman - 2013

1 paper in library cites

[21]Fusing Robust Face Region Descriptors via Multiple Metric Learning for Face Recognition in the Wild

Z. Cui, Wentao Li, D. X. Xu, S. Shan, X. Chen - 2013

1 paper in library cites

[22]Learning Hierarchical Representations for Face Verification With Convolutional Deep Belief Networks

G. B. Huang, Honglak Lee, E. G. L. Miller - 2012

1 paper in library cites

[23]Learning to Align From Scratch

G. B. Huang, M. A. Mattar, Honglak Lee, E. G. L. Miller - 2012

1 paper in library cites

[24]Leveraging Billions of Faces to Overcome Performance Barriers in Unconstrained Face Recognition

Y. Taigman, Lior Wolf - 2011

1 paper in library cites

[25]Multiple One-Shots for Utilizing Class Label Information

Y. Taigman, Lior Wolf, T. Hassner - 2009

1 paper in library cites

[26]Probabilistic Elastic Matching for Pose Variant Face Verification

H. Li, G. Hua, Zongyu Lin, J. Brandt, Jihan Yang - 2013

1 paper in library cites

[27]The SVM-minus Similarity Score for Video Face Recognition

Lior Wolf, N. Levy - 2013

1 paper in library cites

[28]Tom-vs-Pete Classifiers and Identity-Preserving Alignment for Face Verification

T. Berg, P. N. Belhumeur - 2012

1 paper in library cites

[29]Towards Pose Robust Face Recognition

D. Yi, Z. Lei, S. Z. Li - 2013

1 paper in library cites

[30]Unsupervised Joint Alignment of Complex Images

G. B. Huang, V. Jain, E. G. L. Miller - 2007

1 paper in library cites

[31]Viewing Real-World Faces in 3D

T. Hassner - 2013

1 paper in library cites

[32]Volume Structured Ordinal Features With Background Similarity Measure for Video Face Recognition

H. M. Vazquez, Y. M. Diaz, Z. Chai - 2013

1 paper in library cites

Cited by

5

papers in your library

Cites

5

papers in your library

Read

on October 18, 2025

Very impressive results but boring overall. I think the main thing is that they were the first to use CNN, but it seems like the most important part of their method was the 3D alignment (which is nice, but out of scope for me)

Tags

Paper Aliases

No aliases