Eigenfaces vs. Fisherfaces: Recognition Using Class Specific Linear Projection

By Peter N. Belhumeur, João Pedra Hespanha, David J. Kriegman. 1996

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

Accessorize to a Crime: Real and Stealthy Attacks on State-of-the-Art Face Recognition (2016, background): “FengandPrabhakaran proposed a system to help artists come up with makeup and hair designs that can be used to prevent FRSs (including Eigenfaces [40] and Fisherfaces [1]) from recognizing subjects [10].”
Deep Learning Face Representation by Joint Identification-Verification (2014, methodology): “It can be traced back to early subspace face recognition methods such as LDA [1], Bayesian face [17], and unified subspace [23, 24].”
Learning Face Representation from Scratch (2014, background): “Fisherface [1] or LDA is a one layer linear model too.”
2D and 3D face recognition: A survey (2007, methodology): “…Name Martinez and Kak (2001) PCA AR-Faces 85 · 60 100–250 No 70 No No No No Martinez and Kak (2001) LDA AR-Faces 85 · 60 100–250 No 88 No No No No Belhumeur et al. (1997) Fisherfaces YALE 144–16 No 99.6 Yes Yes No No No Yu and Yang (2001) Direct LDA ORL 112 · 92 200–200 No 90.8 Yes Yes Yes…”
Face recognition with radial basis function (RBF) neural networks (2002, methodology): “The optimal subspace, by the FLD is determined as follows [7], [8], [13]:”
The Quotient Image: Class-Based Re-Rendering and Recognition with Varying Illuminations (2001, background): “The class-based recognition problem is similarly defined: given a single image of an object in a database of images of other objects, some of them are multiply sampled under varying illumination, identify (match) any novel image of that object under varying illumination with the single image of…”
Robust Structured Subspace Learning for Data Representation (2015, methodology): “According to the availability of label information, there are three broad categories: supervised [4], [13], [15], semi-supervised [16] and unsupervised feature selection methods [5], [17], [18].”
Optimization Models for Machine Learning: A Survey (2019, background): “1) share the commonality of dimensionality reduction, LDA tends to be more robust than PCA since it takes into account the data labels in computing the optimal projection matrix [18].”
50 years of biometric research: Accomplishments, challenges, and opportunities (2016, unknown): “Consequently , most of the research in biometric recognition has rightly focused on the following two fundamental problems:”
Gabor Feature Based Sparse Representation for Face Recognition with Gabor Occlusion Dictionary (2010, methodology): “Therefore, subspace learning and manifold learning methods have been dominantly and successfully used in appearance based FR [4], [ 5 ], [6], [7], [8], [9], [10], [11].”
Flexible Orthogonal Neighborhood Preserving Embedding (2017, methodology): “Besides, a manifold of data usually exhibits significant non-linear structure while PCA and LDA are both linear dimension reduction methods.”
Random Multispace Quantization as an Analytic Mechanism for BioHashing of Biometric and Random Identity Inputs (2006, methodology): “We found the performance of FDA to be consistently better than that of PCA, which is consistent with the results previously reported in [ 19 ], and that the FDA þ RMQ-m extensions uniformly out-perform PCA þ RMQ-m for corresponding m. Note also that FDA þ RMQ-p and PCA þ RMQ-p analysis yields consis”
A combined SVM and LDA approach for classification (2005, methodology): “Linear Discriminant Analysis Linear discriminant analysis (LDA) has been successfully applied in many applications such as face recognition [1] [12],”