cvpr2008 tutorial generalized pca
DESCRIPTION
part I: 1part II: 12part III: 43part IV: 80part V: 116TRANSCRIPT
![Page 1: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/1.jpg)
Generalized Principal Component AnalysisTutorial @ CVPR 2008
Yi Ma
ECE Department
University of Illinois
Urbana Champaign
René Vidal
Center for Imaging Science
Institute for Computational Medicine
Johns Hopkins University
![Page 2: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/2.jpg)
Data segmentation and clustering
• Given a set of points, separate them into multiple groups
• Discriminative methods: learn boundary
• Generative methods: learn mixture model, using, e.g.Expectation Maximization
![Page 3: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/3.jpg)
Dimensionality reduction and clustering
• In many problems data is high-dimensional: can reducedimensionality using, e.g. Principal Component Analysis
• Image compression
• Recognition– Faces (Eigenfaces)
• Image segmentation– Intensity (black-white)
– Texture
![Page 4: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/4.jpg)
Segmentation problems in dynamic vision
• Segmentation of video and dynamic textures
• Segmentation of rigid-body motions
![Page 5: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/5.jpg)
Segmentation problems in dynamic vision
• Segmentation of rigid-body motions from dynamic textures
![Page 6: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/6.jpg)
Clustering data on non Euclidean spaces
• Clustering data on non Euclidean spaces– Mixtures of linear spaces
– Mixtures of algebraic varieties
– Mixtures of Lie groups
• “Chicken-and-egg” problems– Given segmentation, estimate models
– Given models, segment the data
– Initialization?
• Need to combine– Algebra/geometry, dynamics and statistics
![Page 7: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/7.jpg)
Outline of the tutorial
• Introduction (8.00-8.15)
• Part I: Theory (8.15-9.45)
– Basic GPCA theory and algorithms (8.15-9.00)
– Advanced statistical methods for GPCA (9.00-9.45)
• Questions (9.45-10.00)
• Break (10.00-10.30)
• Part II: Applications (10.30-12.00)
– Applications to motion and video segmentation (10.30-11.15)
– Applications to image representation & segmentation (11.15-12.00)
• Questions (12.00-12.15)
![Page 8: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/8.jpg)
Part I: Theory
• Introduction to GPCA (8.00-8.15)
• Basic GPCA theory and algorithms (8.15-9.00)
– Review of PCA and extensions
– Introductory cases: line, plane and hyperplane segmentation
– Segmentation of a known number of subspaces
– Segmentation of an unknown number of subspaces
• Advanced statistical and methods for GPCA (9.00-9.45)
– Lossy coding of samples from a subspace
– Minimum coding length principle for data segmentation
– Agglomerative lossy coding for subspace clustering
![Page 9: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/9.jpg)
Part II: Applications in computer vision
• Applications to motion & video segmentation (10.30-11.15)
– 2-D and 3-D motion segmentation
– Temporal video segmentation
– Dynamic texture segmentation
• Applications to image representation and segmentation
(11.15-12.00)
– Multi-scale hybrid linear models for sparse
image representation
– Hybrid linear models for image segmentation
![Page 10: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/10.jpg)
References: Springer-Verlag 2008
![Page 11: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/11.jpg)
Slides, MATLAB code, papers
Slides: http://www.vision.jhu.edu/gpca/cvpr08-tutorial-gpca.htm
Code: http://perception.csl.uiuc.edu/gpca
![Page 12: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/12.jpg)
Part I
Generalized Principal Component AnalysisRené Vidal
Center for Imaging Science
Institute for Computational Medicine
Johns Hopkins University
![Page 13: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/13.jpg)
Principal Component Analysis (PCA)
• Given a set of points x1, x2, …, xN
– Geometric PCA: find a subspace S passing through them
– Statistical PCA: find projection directions that maximize the variance
• Solution (Beltrami’1873, Jordan’1874, Hotelling’33, Eckart-Householder-Young’36)
• Applications: data compression, regression, computervision (eigenfaces), pattern recognition, genomics
Basis for S
![Page 14: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/14.jpg)
Extensions of PCA
• Higher order SVD (Tucker’66, Davis’02)
• Independent Component Analysis (Common ‘94)
• Probabilistic PCA (Tipping-Bishop ’99)
– Identify subspace from noisy data
– Gaussian noise: standard PCA
– Noise in exponential family (Collins et al.’01)
• Nonlinear dimensionality reduction
– Multidimensional scaling (Torgerson’58)
– Locally linear embedding (Roweis-Saul ’00)
– Isomap (Tenenbaum ’00)
• Nonlinear PCA (Scholkopf-Smola-Muller ’98)
– Identify nonlinear manifold by applying PCA todata embedded in high-dimensional space
• Principal Curves and Principal Geodesic Analysis(Hastie-Stuetzle’89, Tishbirany ‘92, Fletcher ‘04)
![Page 15: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/15.jpg)
Generalized Principal Component Analysis
• Given a set of points lying in multiple subspaces, identify
– The number of subspaces and their dimensions
– A basis for each subspace
– The segmentation of the data points
• “Chicken-and-egg” problem
– Given segmentation, estimate subspaces
– Given subspaces, segment the data
![Page 16: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/16.jpg)
Prior work on subspace clustering
• Iterative algorithms:
– K-subspace (Ho et al. ’03),
– RANSAC, subspace selection and growing (Leonardis et al. ’02)
• Probabilistic approaches: learn the parameters of a mixture
model using e.g. EM
– Mixtures of PPCA: (Tipping-Bishop ‘99):
– Multi-Stage Learning (Kanatani’04)
• Initialization
– Geometric approaches: 2 planes in R3 (Shizawa-Maze ’91)
– Factorization approaches: independent subspaces of equal
dimension (Boult-Brown ‘91, Costeira-Kanade ‘98, Kanatani ’01)
– Spectral clustering based approaches: (Yan-Pollefeys’06)
![Page 17: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/17.jpg)
Basic ideas behind GPCA
• Towards an analytic solution to subspace clustering
– Can we estimate ALL models simultaneously using ALL data?
– When can we do so analytically? In closed form?
– Is there a formula for the number of models?
• Will consider the most general case
– Subspaces of unknown and possibly different dimensions
– Subspaces may intersect arbitrarily (not only at the origin)
• GPCA is an algebraic geometric approach to data segmentation
– Number of subspaces = degree of a polynomial
– Subspace basis = derivatives of a polynomial
– Subspace clustering is algebraically equivalent to
• Polynomial fitting
• Polynomial differentiation
![Page 18: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/18.jpg)
Applications of GPCA in computer vision
• Geometry
– Vanishing points
• Image compression
• Segmentation
– Intensity (black-white)
– Texture
– Motion (2-D, 3-D)
– Video (host-guest)
• Recognition
– Faces (Eigenfaces)
• Man - Woman
– Human Gaits
– Dynamic Textures
• Water-bird
• Biomedical imaging
• Hybrid systems identification
![Page 19: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/19.jpg)
Introductory example: algebraic clustering in 1D
• Number of groups?
![Page 20: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/20.jpg)
Introductory example: algebraic clustering in 1D
• How to compute n, c, b’s?
– Number of clusters
– Cluster centers
– Solution is unique if
– Solution is closed form if
![Page 21: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/21.jpg)
Introductory example: algebraic clustering in 2D
• What about dimension 2?
• What about higher dimensions?
– Complex numbers in higher dimensions?
– How to find roots of a polynomial of quaternions?
• Instead
– Project data onto one or two dimensional space
– Apply same algorithm to projected data
![Page 22: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/22.jpg)
Representing one subspace
• One plane
• One line
• One subspace can be represented with
– Set of linear equations
– Set of polynomials of degree 1
![Page 23: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/23.jpg)
De Morgan’s rule
Representing n subspaces
• Two planes
• One plane and one line
– Plane:
– Line:
• A union of n subspaces can be represented with a set of
homogeneous polynomials of degree n
![Page 24: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/24.jpg)
Veronese map
Fitting polynomials to data points
• Polynomials can be written linearly in terms of the vector of coefficientsby using polynomial embedding
• Coefficients of the polynomials can be computed from nullspace ofembedded data
– Solve using least squares
– N = #data points
![Page 25: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/25.jpg)
Finding a basis for each subspace
• Case of hyperplanes:
– Only one polynomial
– Number of subspaces
– Basis are normal vectors
• Problems
– Computing roots may be sensitive to noise
– The estimated polynomial may not perfectly factor with noisy
– Cannot be applied to subspaces of different dimensions
• Polynomials are estimated up to change of basis, hence they may not factor,
even with perfect data
Polynomial Factorization (GPCA-PFA) [CVPR 2003]• Find roots of polynomial of degree in one variable
• Solve linear systems in variables
• Solution obtained in closed form for
![Page 26: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/26.jpg)
Finding a basis for each subspace
• To learn a mixture of subspaces we just need one positiveexample per class
Polynomial Differentiation (GPCA-PDA) [CVPR’04]
![Page 27: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/27.jpg)
Choosing one point per subspace
• With noise and outliers
– Polynomials may not be a perfect union of subspaces
– Normals can estimated correctly by choosing points optimally
• Distance to closest subspace without knowing
segmentation?
![Page 28: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/28.jpg)
GPCA for hyperplane segmentation
• Coefficients of the polynomial can be computed from null
space of embedded data matrix
– Solve using least squares
– N = #data points
• Number of subspaces can be computed from the rank of
embedded data matrix
• Normal to the subspaces can be computed
from the derivatives of the polynomial
![Page 29: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/29.jpg)
GPCA for subspaces of different dimensions
• There are multiple polynomials
fitting the data
• The derivative of each
polynomial gives a different
normal vector
• Can obtain a basis for the
subspace by applying PCA to
normal vectors
![Page 30: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/30.jpg)
GPCA for subspaces of different dimensions
• Apply polynomial embedding to projected data
• Obtain multiple subspace model by polynomial fitting
– Solve to obtain
– Need to know number of subspaces
• Obtain bases & dimensions by polynomial differentiation
• Optimally choose one point per subspace using distance
![Page 31: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/31.jpg)
An example
• Given data lying in the union
of the two subspaces
• We can write the union as
• Therefore, the union can be represented with the two
polynomials
![Page 32: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/32.jpg)
An example
• Can compute polynomials from
• Can compute normals from
![Page 33: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/33.jpg)
Dealing with high-dimensional data
• Minimum number of points
– K = dimension of ambient space
– n = number of subspaces
• In practice the dimension ofeach subspace ki is muchsmaller than K
– Number and dimension of thesubspaces is preserved by alinear projection onto asubspace of dimension
– Can remove outliers by robustlyfitting the subspace
• Open problem: how to chooseprojection?
– PCA?
Subspace 1
Subspace 2
![Page 34: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/34.jpg)
GPCA with spectral clustering
• Spectral clustering
– Build a similarity matrix between pairs of points
– Use eigenvectors to cluster data
• How to define a similarity for subspaces?
– Want points in the same subspace to be close
– Want points in different subspace to be far
• Use GPCA to get basis
• Distance: subspace angles
![Page 35: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/35.jpg)
Comparison of PFA, PDA, K-sub, EM
0 1 2 3 4 50
2
4
6
8
10
12
14
16
18
Noise level [%]
Err
or in
the
norm
als
[deg
rees
]PFAK−subPDAEMPDA+K−subPDA+EMPDA+K−sub+EM
![Page 36: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/36.jpg)
Null(Ln) L
n= [
n(x1), . . . ,
n(xN)].
Dealing with outliers
• GPCA with perfect data
• GPCA with outliers
• GPCA fails because PCA fails seek a robust estimateof where
![Page 37: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/37.jpg)
Three approaches to tackle outliers
• Probability-based: small-probability samples
– Probability plots: [Healy 1968, Cox 1968]
– PCs: [Rao 1964, Ganadesikan & Kettenring 1972]
– M-estimators: [Huber 1981, Camplbell 1980]
– Multivariate-trimming (MVT):[Ganadesikan & Kettenring 1972]
• Influence-based: large influence on model parameters
– Parameter difference with and without a sample:[Hampel et al. 1986, Critchley 1985]
• Consensus-based: not consistent with models of high consensus.
– Hough: [Ballard 1981, Lowe 1999]
– RANSAC: [Fischler & Bolles 1981, Torr 1997]
– Least Median Estimate (LME):[Rousseeuw 1984, Steward 1999]
![Page 38: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/38.jpg)
Robust GPCA
![Page 39: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/39.jpg)
Robust GPCA
Simulation on Robust GPCA (parameters fixed at = 0.3rad and = 0.4
• RGPCA – Influence
• RGPCA - MVT
(e) 12% (f) 32% (g) 48% (h) 12% (i) 32% (j) 48%
(k) 12% (l) 32% (m) 48% (n) 12% (o) 32% (p) 48%
![Page 40: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/40.jpg)
Robust GPCA
Comparison with RANSAC
• Accuracy
• Speed
(q) (2,2,1) in 3 (s) (5,5,5) in 6(r) (4,2,2,1) in 5
Arrangement (2,2,1) in 3 (4,2,2,1) in 5 (5,5,5) in 6
RANSAC 44s 5.1min 3.4min
MVT 46s 23min 8min
Influence 3min 58min 146min
Table: Average time of RANSAC and RGPCA with 24% outliers.
![Page 41: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/41.jpg)
Summary
• GPCA: algorithm for clustering subspaces
– Deals with unknown and possibly different dimensions
– Deals with arbitrary intersections among the subspaces
• Our approach is based on
– Projecting data onto a low-dimensional subspace
– Fitting polynomials to projected subspaces
– Differentiating polynomials to obtain a basis
• Applications in image processing and computer vision
– Image segmentation: intensity and texture
– Image compression
– Face recognition under varying illumination
![Page 42: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/42.jpg)
For more information,
Vision, Dynamics and Learning Lab
@
Johns Hopkins University
Thank You!Thank You!
![Page 43: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/43.jpg)
Generalized Principal Component Analysis Generalized Principal Component Analysis via via LossyLossy Coding and CompressionCoding and Compression
Yi MaYi Ma
Image Formation & Processing Group, BeckmanDecision & Control Group, Coordinated Science
Lab.Electrical & Computer Engineering Department
University of Illinois at Urbana-Champaign
![Page 44: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/44.jpg)
MOTIVATION
PROBLEM FORMULATION AND EXISTING APPROACHES
SEGMENTATION VIA LOSSY DATA COMPRESSION
SIMULATIONS (AND EXPERIMENTS)
CONCLUSIONS AND FUTURE DIRECTIONS
OUTLINE
![Page 45: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/45.jpg)
MOTIVATION – Motion Segmentation in Computer Vision
The “chicken-and-egg” difficulty:– Knowing the segmentation, estimating the motions is easy;– Knowing the motions, segmenting the features is easy.
Goal: Given a sequence of images of multiple moving objects, determine:– 1. the number and types of motions (rigid-body, affine, linear, etc.)
2. the features that belong to the same motion.
A Unified Algebraic Approach to 2D and 3D Motion Segmentation, [Vidal-Ma, ECCV’
QuickTime™ and aCinepak decompressor
are needed to see this picture.
![Page 46: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/46.jpg)
MOTIVATION – Image Segmentation
features
Computer Human
Goal: segment an image into multiple regions with homogeneous texture.
Difficulty: A mixture of models of different dimensions or complexities.
Multiscale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP’
![Page 47: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/47.jpg)
MOTIVATION – Video SegmentationGoal: segmenting a video sequence into segments with “stationary” dynamics
Identification of Hybrid Linear Systems via Subspace Segmentation, [Huang-Wagner-Ma, C
Model: different segments as outputs from different (linear) dynamical systems:
QuickTime™ and aH.264 decompressor
are needed to see this picture.
![Page 48: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/48.jpg)
MOTIVATION – Massive Multivariate Mixed Data
Face database
Hand written digits
Hyperspectral images
Microarrays
Articulate motions
QuickTime™ and aBMP decompressor
are needed to see this picture.
![Page 49: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/49.jpg)
SUBSPACE SEGMENTATION – Problem Formulation
Difficulties:– the dimensions of the subspaces can be different– the data can be corrupted by noise or contaminated by outliers– the number and dimensions of subspaces may be unknown
Assumption: the data are noisy samples from an arrangement of linear subspaces:
noise-free samples noisy samples samples with outliers
![Page 50: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/50.jpg)
SUBSPACE SEGMENTATION – Statistical Approaches
Assume that the data are i.i.d. samples from a mixture of probabilistic distributions:
Essentially iterate between data segmentation and model estimation.
Solutions:• Expectation Maximization (EM) for the maximum-likelihood estimate
[Dempster et. al.’77], e.g., Probabilistic PCA [Tipping-Bishop’99]:
• K-Means for a minimax-like estimate [Forgy’65, Jancey’66, MacQueen’67], e.g., K-Subspaces [Ho and Kriegman’03]:
![Page 51: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/51.jpg)
SUBSPACE SEGMENTATION – An Algebro-Geometric Approach
Idea: a union of linear subspaces is an algebraic set -- the zero set of a set of (homogeneous) polynomials:
Complexity exponential in the dimension and number of subspaces.
Solution:• Identify the set of polynomials of degree n that vanish on
• Gradients of the vanishing polynomials are normals to the subspaces
Generalized Principal Component Analysis, [Vidal-Ma-Sastry, IEEE Transactions PAMI’0
![Page 52: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/52.jpg)
SUBSPACE SEGMENTATION – An Information-Theoretic Approach
Problem: If the number/dimension of subspaces not given and data corrupted
by noise and outliers, how to determine the optimal subspaces that fit the data?Solutions: Model Selection Criteria?– Minimum message length (MML) [Wallace-Boulton’68]– Minimum description length (MDL) [Rissanen’78]– Bayesian information criterion (BIC)– Akaike information criterion (AIC) [Akaike’77]– Geometric AIC [Kanatani’03], Robust AIC [Torr’98]
Key idea (MDL):• a good balance between model complexity and data fidelity.• minimize the length of codes that describe the model and the data:
with a quantization error optimal for the model.
![Page 53: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/53.jpg)
LOSSY DATA COMPRESSION
Questions:
– What is the “gain” or “loss” of segmenting or merging data?
– How does tolerance of error affect segmentation results?
Basic idea: whether the number of bits required to store “the whole is more than the sum of its parts”?
![Page 54: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/54.jpg)
LOSSY DATA COMPRESSION – Problem Formulation
– A coding scheme maps a set of vectors to a sequence of bits, from which we can decode The coding length is denoted as:
– Given a set of real-valued mixed data the optimal segmentation
minimizes the overall coding length:
where
![Page 55: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/55.jpg)
LOSSY DATA COMPRESSION – Coding Length for Multivariate Data
Theorem.Given with
is the number of bits needed to encode the data s.t. .
A nearly optimal bound for even a small number of vectors drawn from a subspace or a Gaussian source.
Segmentation of Multivariate Mixed Data, [Ma-Derksen-Hong-Wright, PAMI’
![Page 56: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/56.jpg)
LOSSY DATA COMPRESSION – Two Coding Schemes
Goal: code s.t. a mean squared error
Linear subspace Gaussian source
![Page 57: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/57.jpg)
LOSSY DATA COMPRESSION – Properties of the Coding Length
2. Asymptotic Property:
At high SNR, this is the optimal rate distortion for a Gaussian source.
1. Commutative Property:For high-dimensional data, computing the coding length only needsthe kernel matrix:
3. Invariant Property:Harmonic Analysis is useful for data compression only when the data arenon-Gaussian or nonlinear ……… so is segmentation!
![Page 58: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/58.jpg)
LOSSY DATA COMPRESSION – Why Segment?
partitioning:
sifting:
![Page 59: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/59.jpg)
LOSSY DATA COMPRESSION – Probabilistic Segmentation?
is a concave function in Π over the domain of a convex polytope.
Minima are reached at the vertexes of the polytope -- no probabilistic segmentation!
Assign the ith point to the jth group with probability
Theorem. The expected coding length of the segmented data
Segmentation of Multivariate Mixed Data, [Ma-Derksen-Hong-Wright, PAMI’
![Page 60: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/60.jpg)
LOSSY DATA COMPRESSION – Segmentation & Channel Capacity
A MIMO additive white Gaussian noise (AWGN) channel
has the capacity:
If allowing probabilistic grouping of transmitters, the expectedcapacity
is a concave function in Π over a convex polytope.Maximizing such a capacity is a convexproblem.
On Coding and Segmentation of Multivariate Mixed Data, [Ma-Derksen-Hong-Wright, PAMI
![Page 61: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/61.jpg)
LOSSY DATA COMPRESSION – A Greedy (Agglomerative) Algorithm
Objective: minimizing the overall coding length
Input:
while true dochoose two sets such
that is minimal
ifthenelse breakendif
endOutput:
“Bottom-up” merge
QuickTime™ and aPNG decompressor
are needed to see this picture.
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
![Page 62: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/62.jpg)
SIMULATIONS – Mixture of Almost Degenerate GaussiansNoisy samples from two lines and one plane in <3
Given Data Segmentation Results
ε0 = 0.01
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
ε0 = 0.08
![Page 63: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/63.jpg)
SIMULATIONS – “Phase Transition”Rate v.s. distortion
0.08
ε0 = 0.08
#group v.s. distortion
Stability: the same segmentation for ε across 3 magnitudes!
0.08
steamwater
ice cubes
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
![Page 64: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/64.jpg)
SIMULATIONS – Comparison with EM
100 x d uniformly distributed random samples from each subspace, corruptewith 4% noise. Classification rate averaged over 25 trials for each case.
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
![Page 65: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/65.jpg)
SIMULATIONS – Comparison with EM
Segmenting three degenerate or non-degenerate Gaussian clusters for 50 tria
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
![Page 66: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/66.jpg)
SIMULATIONS – Robustness with Outliers
35.8% outliers 45.6%
71.5% 73.6%
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
![Page 67: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/67.jpg)
SIMULATIONS – Affine Subspaces with Outliers
35.8% outliers 45.6%
66.2% 69.1%
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
![Page 68: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/68.jpg)
SIMULATIONS – Piecewise-Linear Approximation of Manifolds
Swiss roll Mobius strip Torus Klein bottle
![Page 69: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/69.jpg)
SIMULATIONS – Summary
– The minimum coding length objective automatically addresses the
model selection issue: the optimal solution is very stable and robust.
– The segmentation/merging is physically meaningful (measured in bits).
The results resemble phase transition in statistical physics.
– The greedy algorithm is scalable (polynomial in both K and N) and
converges well when ε is not too small w.r.t. the sample density.
![Page 70: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/70.jpg)
Clustering from a Classification Perspective
Solution: Knowing the distributions and , the optimal classifier is the maximum a posteriori (MAP) classifier:
Difficulties: How to learn the two distributions from samples?(parametric, non-parametric, model selection, high-dimension, outliers…)
Goal: Construct a classifier such that the misclassification error
reaches minimum.
Assumption: The training data are drawn from a distribution
![Page 71: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/71.jpg)
MINIMUM INCREMENTAL CODING LENGTH – Problem Formulation
Ideas: Using the lossy coding length
as a surrogate for the Shannon lossless coding length w.r.t. true distributions.
Classification Criterion: Minimum Incremental Coding Length (MICL)
Additional bits need to encode the test sample with the jth training set is
![Page 72: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/72.jpg)
MICL (“Michael”) – Asymptotic Properties
Theorem: As the number of samples goes to infinity, the MICL criterion converges with probability one to the following criterion:
where?
is the “number of effective parameters” of the j-th model (class).
Theorem: The MICL classifier converges to the above asymptotic form at the rate of for some constant .
Minimum Incremental Coding Length (MICL), [Wright and Ma et. a.., NIPS’07]
![Page 73: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/73.jpg)
SIMULATIONS – Interpolation and Extrapolation via MICL
MICL
SVM
k-NN
Minimum Incremental Coding Length (MICL), [Wright and Ma et. a.., NIPS’07]
![Page 74: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/74.jpg)
SIMULATIONS – Improvement over MAP and RDA [Friedman1989]
Two Gaussians in R2
isotropic (left)anisotropic
(right)(500 trials)
Three Gaussians in Rn
dim = ndim = n/2dim = 1
(500 trials)
Minimum Incremental Coding Length (MICL), [Wright and Ma et. a.., NIPS’07]
![Page 75: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/75.jpg)
SIMULATIONS – Local and Kernel MICL
LMICL k-NN
KMICL-RBF SVM-RBF
Local MICL (LMICL): Applying MICL locally to the k-nearest neighbors of the test sample (frequencylist + Bayesianist).
Kernel MICL (KMICL): Incorporating MICL with a nonlinear kernel naturally through the identity (“kernelized” RDA):
Minimum Incremental Coding Length (MICL), [Wright and Ma et. a.., NIPS’07]
![Page 76: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/76.jpg)
CONCLUSIONS
Assumptions: Data are in a high-dimensional space but have low-dimensional structures (subspaces or submanifolds).
Compression => Clustering & Classification:– Minimum (incremental) coding length subject to distortion.– Asymptotically optimal clustering and classification.– Greedy clustering algorithm (bottom-up, agglomerative).– MICL corroborates MAP, RDA, k-NN, and kernel methods.
Applications (Next Lectures):– Video segmentation, motion segmentation (Vidal)– Image representation & segmentation (Ma)– Others: microarray clustering, recognition of faces and
handwritten digits (Ma)
![Page 77: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/77.jpg)
FUTURE DIRECTIONS
Theory– More complex structures: manifolds, systems, random
fields…– Regularization (ridge, lasso, banding etc.)– Sparse representation and subspace arrangements
Computation– Global optimality (random techniques, convex
optimization…)– Scalability: random sampling, approximation…
Future Application Domains– Image/video/audio classification, indexing, and retrieval– Hyper-spectral images and videos– Biomedical images, microarrays– Autonomous navigation, surveillance, and 3D mapping– Identification of hybrid linear/nonlinear systems
![Page 78: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/78.jpg)
REFERENCES & ACKNOWLEGMENT
References:– Segmentation of Multivariate Mixed Data via Lossy Data
Compression, Yi Ma, Harm Derksen, Wei Hong, John Wright, PAMI, 2007.
– Classification via Minimum Incremental Coding Length (MICL), John Wright et. al., NIPS, 2007.
– Website: http://perception.csl.uiuc.edu/coding/home.htm
People:– John Wright, PhD Student, ECE Department, University of Illinois– Prof. Harm Derksen, Mathematics Department, University of
Michigan– Allen Yang (UC Berkeley) and Wei Hong (Texas Instruments R&D)– Zhoucheng Lin and Harry Shum, Microsoft Research Asia, China
Funding:– ONR YIP N00014-05-1-0633– NSF CAREER IIS-0347456, CCF-TF-0514955, CRS-EHS-0509151
![Page 79: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/79.jpg)
““The whole is more than the sum of its The whole is more than the sum of its partsparts.”.”
----Aristotle Aristotle
Questions, please?
11/2003
Yi Ma, CVPR 2008
![Page 80: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/80.jpg)
Part II
Applications of GPCA in Computer VisionRené Vidal
Center for Imaging Science
Institute for Computational Medicine
Johns Hopkins University
![Page 81: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/81.jpg)
Part II: Applications in computer vision
• Applications to motion & video segmentation (10.30-11.15)
– 2-D and 3-D motion segmentation
– Temporal video segmentation
– Dynamic texture segmentation
• Applications to image representation and segmentation
(11.15-12.00)
– Multi-scale hybrid linear models for sparse
image representation
– Hybrid linear models for image segmentation
![Page 82: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/82.jpg)
Applications to motion and and video
segmentationRené Vidal
Center for Imaging Science
Institute for Computational Medicine
Johns Hopkins University
![Page 83: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/83.jpg)
3-D motion segmentation problem
• Given a set of point correspondences in multiple views, determine
– Number of motion models
– Motion model: affine, homography, fundamental matrix, trifocal tensor
– Segmentation: model to which each pixel belongs
• Mathematics of the problem depends on
– Number of frames (2, 3, multiple)
– Projection model (affine, perspective)
– Motion model (affine, translational, homography, fundamental matrix, etc.)
– 3-D structure (planar or not)
![Page 84: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/84.jpg)
Taxonomy of problems
• 2-D Layered representation– Probabilistic approaches: Jepson-Black’93, Ayer-Sawhney’95, Darrel-Pentland’95, Weiss-
Adelson’96, Weiss’97, Torr-Szeliski-Anandan’99
– Variational approaches: Cremers-Soatto ICCV’03
– Initialization: Wang-Adelson’94, Irani-Peleg’92, Shi-Malik‘98, Vidal-Singaraju’05-’06
• Multiple rigid motions in two perspective views– Probabilistic approaches: Feng-Perona’98, Torr’98
– Particular cases: Izawa-Mase’92, Shashua-Levin’01, Sturm’02,
– Multibody fundamental matrix: Wolf-Shashua CVPR’01, Vidal et al. ECCV’02, CVPR’03, IJCV’06
– Motions of different types: Vidal-Ma-ECCV’04, Rao-Ma-ICCV’05
• Multiple rigid motions in three perspective views– Multibody trifocal tensor: Hartley-Vidal-CVPR’04
• Multiple rigid motions in multiple affine views– Factorization-based: Costeira-Kanade’98, Gear’98, Wu et al.’01, Kanatani’ et al.’01-02-04
– Algebraic: Yan-Pollefeys-ECCV’06, Vidal-Hartley-CVPR’04
• Multiple rigid motions in multiple perspective views
– Schindler et al. ECCV’06, Li et al. CVPR’07
![Page 85: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/85.jpg)
A unified approach to motion segmentation
• Estimation of multiple motion models equivalent to
estimation of one multibody motion model
– Eliminate feature clustering: multiplication
– Estimate a single multibody motion model: polynomial fitting
– Segment multibody motion model: polynomial differentiation
chicken-and-egg
![Page 86: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/86.jpg)
A unified approach to motion segmentation
• Applies to most motion models in computer vision
• All motion models can be segmented algebraically by– Fitting multibody model: real or complex polynomial to all data
– Fitting individual model: differentiate polynomial at a data point
![Page 87: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/87.jpg)
Segmentation of 3-D translational motions
• Multiple epipoles (translation)
• Epipolar constraint: plane in
– Plane normal = epipoles
– Data = epipolar lines
• Multibody epipolar constraint • Epipoles are derivatives of
at epipolar lines
![Page 88: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/88.jpg)
Segmentation of 3-D translational motions
![Page 89: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/89.jpg)
Single-body factorization
• Affine camera model
– p = point
– f = frame
• Motion of one rigid-body lives in a 4-D subspace(Boult and Brown ’91,
Tomasi and Kanade ‘92)
– P = #points
– F = #frames
Structure = 3D surface
Motion = camera position and orientation
![Page 90: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/90.jpg)
Multi-body factorization
• Given n rigid motions
• Motion segmentation is obtained from
– Leading singular vector of (Boult and Brown ’91)
– Shape interaction matrix (Costeira & Kanade ’95, Gear ’94)
– Number of motions (if fully-dimensional)
• Motion subspaces need to be independent (Kanatani ’01)
![Page 91: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/91.jpg)
Multi-body factorization
• Sensitive to noise
– Kanatani (ICCV ’01): use model selection to scale Q
– Wu et al. (CVPR’01): project data onto subspaces and iterate
• Fails with partially dependent motions
– Zelnik-Manor and Irani (CVPR’03)
• Build similarity matrix from normalized Q
• Apply spectral clustering to similarity matrix
– Yan and Pollefeys (ECCV’06)
• Local subspace estimation + spectral clustering
– Kanatani (ECCV’04)
• Assume degeneracy is known: pure translation in the image
• Segment data by multi-stage optimization (multiple EM problems)
• Cannot handle missing data
– Gruber and Weiss (CVPR’04)
• Expectation Maximization
![Page 92: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/92.jpg)
PowerFactorization+GPCA
• A motion segmentation algorithm that
– Is provably correct with perfect data
– Handles both independent and degenerate motions
– Handles both complete and incomplete data
• Project trajectories onto a 5-D subspace of
– Complete data: PCA or SVD
– Incomplete data: PowerFactorization
• Cluster projected subspaces using GPCA
– Handles both independent and degenerate motions
– Non-iterative: can be used to initialize EM
![Page 93: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/93.jpg)
Projection onto a 5-D subspace
• Motion of one rigid-body lives in
4-D subspace of
• By projecting onto a 5-D
subspace of
– Number and dimensions of
subspaces are preserved
– Motion segmentation is
equivalent to clustering
subspaces of dimension
2, 3 or 4 in
– Minimum #frames = 3
(CK needs a minimum of 2n
frames for n motions)
– Can remove outliers by robustly
fitting the 5-D subspace using
Robust SVD (DeLaTorre-Black)
• What projection to use?
– PCA: 5 principal components
– RPCA: with outliers
Motion 1
Motion 2
![Page 94: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/94.jpg)
Projection onto a 5-D subspace
PowerFactorization algorithm:
• Complete data
– Given A solve for B
– Orthonormalize B
– Given B solve for A
– Iterate
• Converges to rank-rapproximation with rate
Given , factor it as
• Incomplete data
• It diverges in some cases
• Works well with up to 30% ofmissing data
Linear problem
![Page 95: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/95.jpg)
Motion segmentation using GPCA
• Apply polynomial embedding to 5-D points
Veronese map
![Page 96: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/96.jpg)
Hopkins 155 motion segmentation database
• Collected 155 sequences– 120 with 2 motions
– 35 with 3 motions
• Types of sequences– Checkerboard sequences: mostly full
dimensional and independent motions
– Traffic sequences: mostly degenerate (linear,planar) and partially dependent motions
– Articulated sequences: mostly full dimensionaland partially dependent motions
• Point correspondences– In few cases, provided by Kanatani & Pollefeys
– In most cases, extracted semi-automaticallywith OpenCV
![Page 97: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/97.jpg)
Experimental results: Hopkins 155 database
• 2 motions, 120 sequences, 266 points, 30 frames
![Page 98: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/98.jpg)
Experimental results: Hopkins 155 database
• 3 motions, 35 sequences, 398 points, 29 frames
![Page 99: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/99.jpg)
Experimental results: missing data sequences
• There is no clear correlation between amount of missing data and
percentage of misclassification
• This could be because convergence of PF depends more on “where”
missing data is located than on “how much” missing data there is
![Page 100: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/100.jpg)
Conclusions
• For two motions
– Algebraic methods (GPCA and LSA) are more accurate than
statistical methods (RANSAC and MSL)
– LSA performs better on full and independent sequences, while
GPCA performs better on degenerate and partially dependent
– LSA is sensitive to dimension of projection: d=4n better than d=5
– MSL is very slow, RANSAC and GPCA are fast
• For three motions
– GPCA is not very accurate, but is very fast
– MSL is the most accurate, but it is very slow
– LSA is almost as accurate as MSL and almost as fast as GPCA
![Page 101: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/101.jpg)
Segmentation of Dynamic Textures
René Vidal
Center for Imaging Science
Institute for Computational Medicine
Johns Hopkins University
![Page 102: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/102.jpg)
Modeling a dynamic texture: fixed boundary
• Examples of dynamic textures:
• Model temporal evolution as the output of a linear
dynamical system (LDS): Soatto et al. ‘01
dynamics
appearance
images
zt+1
= Azt+ v
t
yt= Cz
t+ w
t
![Page 103: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/103.jpg)
Segmenting non-moving dynamic textures
• One dynamic texture lives in the observability subspace
• Multiple textures live in multiple subspaces
• Cluster the data using GPCA
water
steam
zt+1
= Azt+ v
t
yt= Cz
t+ w
t
![Page 104: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/104.jpg)
Segmenting moving dynamic textures
![Page 105: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/105.jpg)
Segmenting moving dynamic textures
Ocean-bird
![Page 106: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/106.jpg)
Level-set intensity-based segmentation
• Chan-Vese energy functional
• Implicit methods– Represent C as the zero level set
of an implicit function , i.e.C = {(x, y) : (x, y) = 0}
• Solution– The solution to the gradient descent algorithm for is given by
– c1 and c2 are the mean intensities inside and outside the contour C.
![Page 107: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/107.jpg)
Dynamics & intensity-based energy
• We represent the intensities of the pixels in the images as
the output of a mixture of AR models of order p
• We propose the following spatial-temporal extension of the
Chan-Vese energy functional
where
![Page 108: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/108.jpg)
Variational segmentation of dynamic textures
• Given the ARX parameters, we can solve for the implicit
function by solving the PDE
• Given the implicit function , we can solve for the ARX
parameters of the jth region by solving the linear system
![Page 109: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/109.jpg)
Variational segmentation of dynamic textures
• Fixed boundary segmentation results and comparison
Ocean-smoke Ocean-dynamics Ocean-appearance
![Page 110: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/110.jpg)
Variational segmentation of dynamic textures
• Moving boundary segmentation results and comparison
Ocean-fire
![Page 111: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/111.jpg)
Variational segmentation of dynamic textures
• Results on a real sequence
Raccoon on River
![Page 112: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/112.jpg)
Temporal video segmentation
• Segmenting N=30 frames of a
sequence containing n=3
scenes
– Host
– Guest
– Both
• Image intensities are output of
linear system
• Apply GPCA to fit n=3
observability subspaces
dynamics
appearanceimages
xt+1=Axt+vtyt=Cxt+wt
![Page 113: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/113.jpg)
Temporal video segmentation
• Segmenting N=60 frames of a
sequence containing n=3
scenes
– Burning wheel
– Burnt car with people
– Burning car
• Image intensities are output of linearsystem
• Apply GPCA to fit n=3 observabilitysubspaces
dynamics
appearanceimages
xt+1=Axt+vtyt=Cxt+wt
![Page 114: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/114.jpg)
Conclusions
• Many problems in computer vision can be posed as subspaceclustering problems
– Temporal video segmentation
– 2-D and 3-D motion segmentation
– Dynamic texture segmentation
– Nonrigid motion segmentation
• These problems can be solved using GPCA: an algorithm for clusteringsubspaces
– Deals with unknown and possibly different dimensions
– Deals with arbitrary intersections among the subspaces
• GPCA is based on
– Projecting data onto a low-dimensional subspace
– Recursively fitting polynomials to projected subspaces
– Differentiating polynomials to obtain a basis
![Page 115: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/115.jpg)
For more information,
Vision, Dynamics and Learning Lab
@
Johns Hopkins University
Thank You!Thank You!
![Page 116: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/116.jpg)
Generalized Principal Component Analysis Generalized Principal Component Analysis for Image Representation & Segmentationfor Image Representation & Segmentation
Yi MaYi MaControl & Decision, Coordinated Science Laboratory
Image Formation & Processing Group, BeckmanDepartment of Electrical & Computer Engineering
University of Illinois at Urbana-Champaign
![Page 117: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/117.jpg)
INTRODUCTION
GPCA FOR LOSSY IMAGE REPRESENTATION
IMAGE SEGMENTATION VIA LOSSY COMPRESSION
OTHER APPLICATIONS
CONCLUSIONS AND FUTURE DIRECTIONS
![Page 118: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/118.jpg)
Introduction – Image Representation via Linear Transformations
better representations?
pixel-based representation
three matrixes of RGB-values
linear transformationa more compact representation
![Page 119: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/119.jpg)
Introduction
Fixed Orthogonal Bases (representation, approximation, compression) - Discrete Fourier transform (DFT) or discrete cosine transform (DCT)
(Ahmed ’74): JPEG.- Wavelets (multi-resolution) (Daubechies’88, Mallat’92): JPEG-2000.- Curvelets and contourlets (Candes & Donoho’99, Do & Veterlli’00)
Unorthogonal Bases (for redundant representations)- Extended lapped transforms, frames, sparse representations (Lp
geometry)…
Discrete Fourier transform (DFT)
Wavelet transform
6.25% coefficients.
![Page 120: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/120.jpg)
Introduction
Adaptive Bases (optimal if imagery data are uni-modal)- Karhunen-Loeve transform (KLT), also known as PCA (Pearson’1901,
Hotelling’33, Jolliffe’86)
stack
adaptive bases
![Page 121: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/121.jpg)
Introduction – Principal Component Analysis (PCA)Dimensionality Reduction
Find a low-dimensional representation (model) for high-dimensional data.
Variations of PCA – Nonlinear Kernel PCA (Scholkopf-Smola-Muller’98) – Probabilistic PCA (Tipping-Bishop’99, Collins et.al’01)– Higher-Order SVD (HOSVD) (Tucker’66, Davis’02)– Independent Component Analysis (Hyvarinen-Karhunen-Oja’01)
Principal Component Analysis (Pearson’1901, Hotelling’1933, Eckart & Young’1936) or Karhunen-Loeve transform (KLT).
Basis for S SVD
![Page 122: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/122.jpg)
Distribution of the first three principal components of the Baboon image: A clear multi-modal distribution
Hybrid Linear Models – Multi-Modal Characteristics
![Page 123: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/123.jpg)
Vector Quantization (VQ)
- multiple 0-dimensional affine subspaces (i.e. cluster means)
- existing clustering algorithms are iterative (EM, K-means)
Hybrid Linear Models – Multi-Modal Characteristics
![Page 124: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/124.jpg)
Hybrid Linear Models – Versus Linear ModelsA single linear model
stackLinear
Hybrid linear
Hybrid linear models
stack
![Page 125: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/125.jpg)
Hybrid Linear Models – Characteristics of Natural Images
Multivariate Hybrid Hierarchical High-dimensio1D 2D (multi-modal) (multi-scale) (vector-value
Fourier (DCT)
WaveletsCurvelets
Random fields
PCA/KLT
VQ
Hybrid linear
We need a new & simple paradigm to effectively account for allthese characteristics simultaneously.
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
![Page 126: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/126.jpg)
Hybrid Linear Models – Subspace Estimation and Segmentation
“Chicken-and-Egg” Coupling– Given segmentation, estimate subspaces– Given subspaces, segment the data
Hybrid Linear Models (or Subspace Arrangements)
– the number of subspaces is unknown– the dimensions of the subspaces are unknown– the basis of the subspaces are unknown– the segmentation of the data points is unknown
![Page 127: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/127.jpg)
Hybrid Linear Models – Recursive GPCA (an Example)
![Page 128: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/128.jpg)
Hybrid Linear Models – Effective Dimension
Model Selection (for Noisy Data)Model complexity;Data fidelity;
Dimension of each
subspace
Number of points in each
subspace
Number of subspaces
Total number of
points
Model selection criterion: minimizing effective dimension subject to a given error tolerance (or PSNR)
![Page 129: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/129.jpg)
Hybrid Linear Models – Simulation Results (5% Noise)
ED=3
ED=2.0067
ED=1.6717
![Page 130: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/130.jpg)
Hybrid Linear Models – Subspaces of the Barbara Image
![Page 131: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/131.jpg)
PCA (8x8)
HarrWavelet GPCA (8x8)
Original
Hybrid Linear Models – Lossy Image Representation (Baboon)
DCT (JPEG)
GPCA
![Page 132: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/132.jpg)
Multi-Scale Implementation – Algorithm Diagram
Diagram for a level-3 implementation of hybrid linear modelsfor image representation
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 133: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/133.jpg)
Multi-Scale Implementation – The Baboon Image
downsampleby two twice
segmentation of2 by 2 blocks
The Baboon image
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 134: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/134.jpg)
Multi-Scale Implementation – Comparison with Other Methods
The Baboon image
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 135: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/135.jpg)
Multi-Scale Implementation – Image Approximation
Comparison with level-3 wavelet (7.5% coefficients)
Level-3 bior-4.4 wavelets PSNR=23.94
Level-3 hybrid linear model PSNR=24.64
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 136: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/136.jpg)
Multi-Scale Implementation – Block Size Effect
Some problems with the multi-scale hybrid linear model: 1. has minor block effect; 2. is computationally more costly (than Fourier, wavelets, PCA); 3. does not fully exploit spatial smoothness as wavelets.
The Baboon image
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 137: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/137.jpg)
Multi-Scale Implementation – The Wavelet Domain
The Baboon image
segmentationat each scale
LH HH
HL
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 138: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/138.jpg)
Multi-Scale Implementation – Wavelets v.s. Hybrid Linear Wavelets
The Baboon image
Advantages of the hybrid linear model in wavelet domain: 1. eliminates block effect; 2. is computationally less costly (than in the spatial domain); 3. achieves higher PSNR.
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 139: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/139.jpg)
Multi-Scale Implementation – Visual ComparisonComparison among several models (7.5% coefficients)
Original Image
Wavelets PSNR=23.94
Hybrid modelin spatial domain
PSNR=24.64
Hybrid modelin wavelet
domain PSNR=24.88
Multi-Scale Hybrid Linear Models for Lossy Image Representation, [Hong-Wright-Ma, TIP
![Page 140: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/140.jpg)
Image Segmentation – via Lossy Data Compression
stack
QuickTime™ and aPNG decompressor
are needed to see this picture.
![Page 141: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/141.jpg)
APPLICATIONS – Texture-Based Image Segmentation
Naïve approach:– Take a 7x7 Gaussian window around every pixel.– Stack these windows as vectors.– Clustering the vectors using our algorithm.
A few results:
Segmentation of Multivariate Mixed Data via Lossy Coding and Compression, [Ma-Derksen-Hong-Wright, PAMI’07]
![Page 142: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/142.jpg)
APPLICATIONS – Distribution of Texture Features
Question: why does such a simple algorithm work at all?
Answer: Compression (MDL/MCL) is well suited to mid-level texture segmentation.
Using a single representation (e.g. windows, filterbank responses) for texturesdifferent complexity ⇒ redundancy and degeneracy, which can be exploited foclustering / compression.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
Above: singular values of feature vectors from two differentsegments of the image at left.
![Page 143: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/143.jpg)
APPLICATIONS – Compression-based Texture Merging (CTM)
Problem with the naïve approach:Strong edges, segment boundaries
Solution:Low-level, edge-preserving over-segmentation into small homogeneous regions.
Simple features: stacked Gaussian windows (7x7 in our experiments).
Merge adjacent regions to minimize coding length (“compress” the features).
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
QuickTime™ and aTIFF (LZW) decompressor
are needed to see this picture.
![Page 144: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/144.jpg)
APPLICATIONS – Hierarchical Image Segmentation via CTM
ε = 0.1 ε = 0.2 ε = 0.4Lossy coding with varying distortion ε => hierarchy of segmentations
![Page 145: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/145.jpg)
APPLICATIONS – CTM: Qualitative Results
![Page 146: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/146.jpg)
APPLICATIONS – CTM: Quantitative Evaluation and Comparison
PRI: Probabilistic Rand Index [Pantofaru 2005]VoI: Variation of Information [Meila 2005]GCE: Global Consistency Error [Martin 2001]BDE: Boundary Displacement Error [Freixenet 2002]
Berkeley Image Segmentation Database
Unsupervised Segmentation of Natural Images via Lossy Data Compression, CVIU, 200
![Page 147: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/147.jpg)
Other Applications: Multiple Motion Segmentation (on Hopkins155)
Shankar Rao, Roberton Tron, Rene Vidal, and Yi Ma, to appear in CVPR’08
QuickTime™ and aCinepak decompressor
are needed to see this picture.
QuickTime™ and aCinepak decompressor
are needed to see this picture.
Two Motions: MSL 4.14%, LSA 3.45%, ALC 2.40%, and work with up to 25% outliers.Three Motions: MSL 8.32%, LSA 9.73%, ALC 6.26%.
![Page 148: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/148.jpg)
Other Applications – Clustering of Microarray Data
Segmentation of Multivariate Mixed Data, [Ma-Derksen-Hong-Wright, PAMI’
![Page 149: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/149.jpg)
Other Applications – Clustering of Microarray Data
Segmentation of Multivariate Mixed Data, [Ma-Derksen-Hong-Wright, PAMI’
![Page 150: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/150.jpg)
Other Applications – Supervised Classification
Premises: Data lie on anarrangement of subspaces
Unsupervised Clustering– Generalized PCA
Supervised Classification– Sparse Representation
![Page 151: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/151.jpg)
Other Applications – Robust Face Recognition
Robust Face Recognition via Sparse Representation, to appear in PAMI 2008
![Page 152: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/152.jpg)
Other Applications: Robust Motion Segmentation (on Hopkins155)
Shankar Rao, Roberto Tron, Rene Vidal, and Yi Ma, to appear in CVPR’08
Dealing with incomplete or mistracked features with dataset 80% corrupted!
![Page 153: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/153.jpg)
Three Measures of Sparsity: Bits, L_0 and L1-Norm
Reason: High-dimensional data, like images, do have compact, compressible, sparse structures, in terms of their geometry, statistics, and semantics.
![Page 154: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/154.jpg)
Conclusions
Most imagery data are high-dimensional, statistically or geometrically heterogeneous, and have multi-scalestructures.
Imagery data require hybrid models that can adaptivelyrepresent different subsets of the data with different (sparse) linear models.
Mathematically, it is possible to estimate and segment hybrid (linear) models non-iteratively. GPCA offers one such method.
Hybrid models lead to new paradigms, new principles, andnew applications for image representation, compression, and segmentation.
![Page 155: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/155.jpg)
Future Directions
Mathematical Theory– Subspace arrangements (algebraic properties).– Extension of GPCA to more complex algebraic varieties (e.g.,
hybrid multilinear, high-order tensors).– Representation & approximation of vector-valued functions.
Computation & Algorithm Development– Efficiency, noise sensitivity, outlier elimination.– Other ways to combine with wavelets and curvelets.
Applications to Other Data– Medical imaging (ultra-sonic, MRI, diffusion tensor…)– Satellite hyper-spectral imaging.– Audio, video, faces, and digits.– Sensor networks (location, temperature, pressure, RFID…)– Bioinformatics (gene expression data…)
![Page 156: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/156.jpg)
Acknowledgement
People– Wei Hong, Allen Yang, John Wright, University of Illinois– Rene Vidal of Biomedical Engineering Dept., Johns Hopkins
University– Kun Huang of Biomedical & Informatics Science Dept., Ohio-
State University
Funding– Research Board, University of Illinois at Urbana-Champaign– National Science Foundation (NSF CAREER IIS-0347456)– Office of Naval Research (ONR YIP N000140510633)– National Science Foundation (NSF CRS-EHS0509151)– National Science Foundation (NSF CCF-TF0514955)
![Page 157: CVPR2008 tutorial generalized pca](https://reader033.vdocuments.mx/reader033/viewer/2022052321/5494da1fac7959182e8b4c7e/html5/thumbnails/157.jpg)
Thank You!Thank You!
Generalized Principal Component Analysis: Generalized Principal Component Analysis: Modeling and Segmentation of Multivariate Mixed Modeling and Segmentation of Multivariate Mixed DataData
Rene Vidal, Yi Ma, and Rene Vidal, Yi Ma, and ShankarShankar SastrySastrySpringerSpringer--VerlagVerlag, to appear, to appear
Yi Ma, CVPR 2008