13.4.1.5 Surveys, Comparisons, Evaluations, Principal Components

Chapter Contents (Back)
PCA. Principal Components. Evaluation, Principal Components. Survey, PCA. Survey, Pricnipal Components.

Gerbrands, J.J.[Jan J.],
On the relationships between SVD, KLT and PCA,
PR(14), No. 1-6, 1981, pp. 375-381.
WWW Version. 0309
BibRef

Zhao, L.[Li], Yang, Y.H.[Yee-Hong],
Theoretical analysis of illumination in PCA-based vision systems,
PR(32), No. 4, April 1999, pp. 547-564.
WWW Version. BibRef 9904

Martínez, A.M.[Aleix M.], Kak, A.C.[Avinash C.],
PCA versus LDA,
PAMI(23), No. 2, February 2001, pp. 228-233.
IEEE Abstract. IEEE Top Reference.
WWW Version. 0102
When the training set is small, PCA (Principal Components Analysis) outperforms LDA (Linear Discriminant Analysis) and is less sensitive to different training sets. Applied to faces with occlusions. BibRef

Ramamoorthi, R.[Ravi],
Analytic PCA Construction for Theoretical Analysis of Lighting Variability in Images of a Lambertian Object,
PAMI(24), No. 10, October 2002, pp. 1322-1333.
IEEE Abstract. IEEE Top Reference. 0210
Analysis of the process by changing lighting. BibRef

Guillamet, D., Vitriŕ, J.,
Evaluation of distance metrics for recognition based on non-negative matrix factorization,
PRL(24), No. 9-10, June 2003, pp. 1599-1605.
WWW Version. 0304
BibRef
Earlier:
Determining a suitable metric when using non-negative matrix factorization,
ICPR02(II: 128-131).
IEEE DOI Link 0211
BibRef

Guillamet, D., Vitriŕ, J., Schiele, B.,
Introducing a weighted non-negative matrix factorization for image classification,
PRL(24), No. 14, October 2003, pp. 2447-2454.
WWW Version. 0307
BibRef
Earlier: A1, A3, A2:
Analyzing non-negative matrix factorization for image classification,
ICPR02(II: 116-119).
IEEE DOI Link 0211
BibRef

Guillamet, D., Vitria, J.,
Discriminant basis for object classification,
CIAP01(256-261).
IEEE Top Reference. 0210
BibRef

Guillamet, D.[David], Bressan, M.[Marco], and Vitriŕ, J.[Jordi],
A Weighted Non-negative Matrix Factorization for Local Representations,
CVPR01(I:942-947).
IEEE Abstract. IEEE Top Reference. 0110
Deal with problems of the original formulation to get better representations. BibRef

Bressan, M.[Marco], Guillamet, D.[David], and Vitriŕ, J.[Jordi],
Using an ICA Representation of High Dimensional Data for Object Recognition and Classification,
CVPR01(I:1004-1009).
IEEE Abstract. IEEE Top Reference. 0110
BibRef

Bressan, M.[Marco], Vitriŕ, J.[Jordi],
Independent Modes of Variation in Point Distribution Models,
VF01(123 ff.).
HTML Version. 0209
BibRef

Guillamet, D., Moghaddam, B., Vitria, J.,
Higher-order dependencies in local appearance models,
ICIP03(I: 213-216).
IEEE Abstract. IEEE Top Reference. 0312
BibRef
Earlier: A2, A1, A3:
Local appearance-based models using high-order statistics of image features,
CVPR03(I: 729-735).
IEEE Abstract. IEEE Top Reference. 0307
BibRef

Wang, L.W.[Li-Wei], Wang, X.[Xiao], Zhang, X.R.[Xue-Rong], Feng, J.F.[Ju-Fu],
The equivalence of two-dimensional PCA to line-based PCA,
PRL(26), No. 1, 1 January 2005, pp. 57-60.
WWW Version. 0501
BibRef

Martínez, A.M.[Aleix M.], Zhu, M.L.[Man-Li],
Where Are Linear Feature Extraction Methods Applicable?,
PAMI(27), No. 12, December 2005, pp. 1934-1944.
IEEE DOI Link 0512
Analyze where and why eigen-based linear equations do not work. When the smallest angle between the ith eigenvector and the first i eigenvectors is close to zero, there are problems. BibRef

Gao, H.[Hui], Davis, J.W.[James W.],
Why direct LDA is not equivalent to LDA,
PR(39), No. 5, May 2006, pp. 1002-1006.
WWW Version. 0604
BibRef
Earlier:
Sampling Representative Examples for Dimensionality Reduction and Recognition: Bumping LDA,
ECCV06(III: 275-287).
Springer DOI Link 0608
Linear discriminant analysis; Direct LDA; Small sample size problem BibRef

Chen, P.[Pei], Suter, D.[David],
An Analysis of Linear Subspace Approaches for Computer Vision and Pattern Recognition,
IJCV(68), No. 1, June 2006, pp. 83-106.
Springer DOI Link 0605
Such as PCA or SVD. BibRef

Bethge, M.[Matthias],
Factorial coding of natural images: how effective are linear models in removing higher-order dependencies?,
JOSA-A(23), No. 6, June 2006, pp. 1253-1268.
WWW Version. 0610
BibRef

Vicente, M.A.[M. Asuncion], Hoyer, P.O.[Patrik O.], Hyvarinen, A.[Aapo],
Equivalence of Some Common Linear Feature Extraction Techniques for Appearance-Based Object Recognition Tasks,
PAMI(29), No. 5, May 2007, pp. 896-900.
IEEE DOI Link 0704
Contradictory evaluations of PCA vs. ICA. Whitened PCA may yield identical results to ICA in some cases. Describe the situations where ICA improves on PCA. BibRef

Vicente, M.A.[M. Asunción], Fernández, C.[Cesar], Reinoso, O.[Oscar], Payá, L.[Luis],
3D Object Recognition from Appearance: PCA Versus ICA Approaches,
ICIAR04(I: 547-555).
WWW Version. 0409
BibRef

Gao, Q.X.[Quan-Xue],
Is two-dimensional PCA equivalent to a special case of modular PCA?,
PRL(28), No. 10, 15 July 2007, pp. 1250-1251.
WWW Version. 0706
Modular PCA; Two-dimensional PCA BibRef

Shih, F.Y.[Frank Y.], Zhang, K.[Kai],
A distance-based separator representation for pattern classification,
IVC(26), No. 5, May 2008, pp. 667-672.
WWW Version. 0803
Pattern representation; Classification; Support vector machine; PCA; LDA BibRef

Zheng, W.S.[Wei-Shi], Lai, J.H.[Jian-Huang], Li, S.Z.[Stan Z.],
1D-LDA vs. 2D-LDA: When is vector-based linear discriminant analysis better than matrix-based?,
PR(41), No. 7, July 2008, pp. 2156-2172.
WWW Version. 0804
Fisher's linear discriminant analysis (LDA); Matrix-based representation; Vector-based representation; Pattern recognition BibRef


Chojnacki, W.[Wojciech], van den Hengel, A.J.[Anton J.], Brooks, M.J.[Michael J.],
Generalised Principal Component Analysis: Exploiting Inherent Parameter Constraints,
VISAPP06(217-228).
Springer DOI Link 0711
BibRef

Salgian, A.S.[Andrea Selinger],
Combining local descriptors for 3D object recognition and categorization,
ICPR08(1-4).
IEEE DOI Link 0812
BibRef
Earlier:
Using Multiple Patches for 3D Object Recognition,
BP07(1-6).
IEEE DOI Link 0706
BibRef
Earlier:
Object Recognition Using Local Descriptors: A Comparison,
ISVC06(II: 709-717).
Springer DOI Link 0611
Build on: See also Scale and Affine Invariant Interest Point Detectors. SIFT ( See also Distinctive Image Features from Scale-Invariant Keypoints. ), PCA-SIFT ( See also PCA-SIFT: a more distinctive representation for local image descriptors. ) and keyed context patches ( See also Perceptual Grouping Hierarchy for Appearance-Based 3D Object Recognition, A. ). BibRef

Choksuriwong, A., Laurent, H., Emile, B.,
Comparison of Invariant Descriptors for Object Recognition,
ICIP05(I: 377-380).
IEEE DOI Link 0512
BibRef

Brand, M.,
From Subspace to Submanifold Methods,
BMVC04(xx-yy).
HTML Version. 0508
BibRef

Lenz, R., Bui, T.H.[Thanh Hai],
Recognition of non-negative patterns,
ICPR04(III: 498-501).
IEEE DOI Link 0409
PCA analysis. Prove that the non-negative vauues are right. BibRef

Fortuna, J., Quick, P., Capson, D.W.,
A comparison of subspace methods for accurate position measurement,
Southwest04(16-20).
IEEE Abstract. IEEE Top Reference. 0411
BibRef

Fortuna, J., Schuurman, D.C., Capson, D.W.,
A comparison of PCA and ICA for object recognition under varying illumination,
ICPR02(III: 11-15).
IEEE DOI Link 0211
BibRef

Fortuna, J.[Jeff], Capson, D.W.[David W.],
Improved support vector classification using PCA and ICA feature space modification,
PR(37), No. 6, June 2004, pp. 1117-1129.
WWW Version. 0405
BibRef
And:
ICA filters for lighting invariant face recognition,
ICPR04(I: 334-337).
IEEE DOI Link 0409
BibRef

Wu, Q., Liu, Z., Xiong, Z., Wang, Y., Chen, T., Castleman, K.R.,
On optimal subspaces for appearance-based object recognition,
ICIP02(III: 885-888).
IEEE Abstract. IEEE Top Reference. 0210
BibRef

Fancourt, C.L.[Craig L.], Principe, J.C.[Jose C.],
Soft Competitive Principal Component Analysis Using the Mixture of Experts,
DARPA97(1071-1076). BibRef 9700

Pedersen, F.[Finn], Andersson, L.[Leif], and Bengtsson, E.[Ewert],
Investigating Preprocessing of Multivariate Images in Combination with Principal Component Analysis,
SCIA97(xx-yy) 9705

HTML Version. BibRef

Chapter on Matching and Recognition Using Volumes, High Level Vision Techniques, Invariants continues in
Learning for Principal Components, Eigen Representations .


Last update:Nov 16, 2009 at 19:35:14