The FERET evaluation methodology for face-recognition algorithms

By P. Jonathon Phillips, Hyeonjoon Moon, Syed A. Rizvi, Patrick J. Rauss. 1997

Two of the most critical requirements in support of producing reliable face-recognition systems are a large database of facial images and a testing procedure to evaluate systems. The Face Recognition Technology (FERET) program has addressed both issues through the FERET database of facial images and the establishment of the FERET tests. To date, 14,126 images from 1199 individuals are included in the FERET database, which is divided into development and sequestered portions. In September 1996, the FERET program administered the third in a series of FERET face-recognition tests. The primary objectives of the third test were to (1) assess the state of the art, (2) identify future areas of research, and (3) measure algorithm performance on large databases.

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

AI and the Everything in the Whole Wide World Benchmark (2021, methodology): “The Facial Recognition Technology (FERET) program for automated facial recognition, inaugurated in 1993 [Phillips et al., 2000], was one of the first to do so.”
True Few-Shot Learning with Language Models (2021, methodology): “We evaluate the generalization loss of the algorithm chosen by CV (likewise for MDL): where Reliability is important for many practical or commercial applications where worst-case performance is important, such as image recognition [35, 36], dialogue systems [37, 38], and robotics [39, 40].”
Does Object Recognition Work for Everyone? (2019, background): “For instance, commercial gender classification systems were found to have substantially higher error rates for darker-skinned females than for light-skinned males [4, 34, 36].”
Smart Augmentation Learning an Optimal Data Augmentation Strategy (2017, methodology): “Color FERET [19] Version 2 was collected between December 1993 and August 1996 and made freely available with the intent of promoting the development of face recognition algorithms.”
Learning Face Representation from Scratch (2014, methodology): “is section we review some popular face datasets and representation learning methods. 2.1. Face Dataset Early face datasets were almost collected under controlled environments, such as PIE [22], FERET [21] and so on. Through the efforts of many researchers, we could obtain very high performance on th”
Finding Structure with Randomness: Probabilistic Algorithms for Constructing Approximate Matrix Decompositions (2009, background): “Our next example involves a large, dense matrix derived from the FERET databank of face images [107, 108].”
Face Hallucination: Theory and Practice (2007, methodology): “In our experiments, the high-res faces are collected from public face databases such as AR [27] and FERET [31], and MSRA Cartoon face database [7].”
A review on Gabor wavelets for face recognition (2006, background): “methodology and benchmark were designed to evaluate state of the art face recognition algorithms [38].”
Improved gait recognition by gait dynamics normalization (2006, methodology): “Performance for the identification scenario is captured by the Cumulative Match Characteristic (CMC) [25], which plots identification rates (P I ) within a given rank k.”
Face Recognition Across Non-Uniform Motion Blur, Illumination, and Pose (2015, unknown): “dard and publicly available FERET database [32].”
The CAS-PEAL Large-Scale Chinese Face Database and Baseline Evaluations (2008, background): “has not decreased the identification rate in such a degree as in FERET evaluation [5] and FRVT2000 [6], especially when”
Locally Linear Regression for Pose-Invariant Face Recognition (2007, methodology): “Instead, we select three subsets from the FERET pose database, i.e., “ba” (frontal), “be”(right rotation of 15 ) and “bf”(left rotation of 15 )[30], to form the training set.”
Sparse Coding and Dictionary Learning for Symmetric Positive Definite Matrices: A Kernel Approach (2012, methodology): “We used the ‘b’ subset of the FERET dataset [25], which includes 1400 images from 198 subjects.”
KinectFaceDB: A Kinect Database for Face Recognition (2014, methodology): “, Eigenface [19], Fisherface [20], and local binary patterns (LBP) [21]) report highly accurate results on well controlled datasets, the introduction of new face databases [for example, the face recognition technology (FERET) database [22], the face recognition vendor test (FRVT) database [23], and ”
Demographic Estimation from Face Images: Human vs. Machine Performance (2015, methodology): “Human perception of age Performance measure and accuracy (MAE { , 5-year CS z ) Lanitis et al. (2002) [17] 2D shape, Raw intensity FG-NET (NA, 565) Studied on 32 face images MAE: 4.3 Geng et al. (2007) [19] Manifold of raw intensity YGA (1,600, 8,000) Not studied MAE: 5$6 CS: F: $55%, M: $50% Suo et”
An Image Preprocessing Algorithm for Illumination Invariant Face Recognition (2003, background): “Eigenfaces [19] is a standard benchmark for face recognition algorithms [1].”
Linear discriminant analysis for the small sample size problem: an overview (2014, unknown): “FERET [38] Contains 14,126 images from 1199 individuals.”