Globally Maximizing, Locally Minimizing: Unsupervised Discriminant Projection with Applications to Face and Palm Biometrics

By Jian Yang, David Dapeng Zhang, Jing-yu Yang, Ben Niu. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007

381 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/831cbffbfe39a059b1212d49e8fdfd458d1d01c5

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 Signal Processing

Citations in the corpus

A General Exponential Framework for Dimensionality Reduction (2014, methodology): “Observing LPP, UDP and MFA, the criterion functions of many dimensionality reduction algorithms can be summarized as follows, arg max W or min W tr W T S 2 W −1 W T S 1 W (33) where S 1 and S 2 differ according to different algorithms.”
Transfer Learning via Learning to Transfer (2018, methodology): “Though limited parameter-based algorithms (Yang et al., 2007a; Tommasi et al., 2014) can transfer across domains in heterogeneous label spaces, they can only handle binary classiﬁcation problems.”
Face Recognition with Convolutional Neural Networks and subspace learning (2017, methodology): “Holistic features consist of independent component analysis (ICA) [2], principal component analysis (PCA or Eigenfaces) [3], linear discriminate analysis (LDA or Fisherfaces) [4] , and manifold-related methods such as LPP [5].”
Metaface learning for sparse representation based face recognition (2010, methodology): “For a test sample my from this class, intuitively, y could be well approximated by the linear combination of the samples within Ai, i.e. , ,1 in i j i j i ij Ay s , where 1601978-1-4244-7994-8/10/$26.00 ©2010 IEEE ICIP 2010 ,1 ,2 ,[ , ,..., ] ii nT i i i ni are the coefficients.”
Maximum Neighborhood Margin Discriminant Projection for Classification (2014, background): “The 𝑘 -neighborhood parameter denoted by 𝑊𝑘 for building the weight matrix in LPP and UDP is selected as 𝑊𝑘 = 𝑙 − 1 [21].”
A survey of palmprint recognition (2009, background): “Some researchers have developed new subspace approaches and examined them on palmprints [71-73, 102, 105].”
Comments on "Globally Maximizing, Locally Minimizing: Unsupervised Discriminant Projection with Application to Face and Palm Biometrics" (2007, background): “Note that the experiments in [1] mainly focus on the special cases with small (or nonrepresentative) training data set, where the algorithms have to learn from undersampled distributions.”
Manifold Discriminant Analysis (2009, methodology): “Recently, several algorithms have been developed to provide mappings for the whole data space, e.g., Locality Preserving Projections (LPP) [12], Unsupervised Discriminant Projection (UDP) [30].”
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].”