Download presentation
Presentation is loading. Please wait.
Published byAgnes Jordan Modified over 8 years ago
1
Point Distribution Models Active Appearance Models Compilation based on: Dhruv Batra ECE CMU Tim Cootes Machester
2
Essence of the Idea (cont.) Explain a new example in terms of the model parameters
3
So what’s a model Model “Shape” “texture”
4
Active Shape Models training set
5
Texture Models warp to mean shape
6
Intensity Normalisation Allow for global lighting variations Common linear approach Shift and scale so that Mean of elements is zero Variance of elements is 1 Alternative non-linear approach Histogram equalization Transforms so similar numbers of each grey- scale value
7
Shape: Review of Construction Mark face region on training set Sample region Normalise Statistical Analysis The Fun Step
8
Multivariate Statistical Analysis Need to model the distribution of normalised vectors Generate plausible new examples Test if new region similar to training set Classify region
9
Fitting a gaussian Mean and covariance matrix of data define a gaussian model
10
Principal Component Analysis Compute eigenvectors of covariance, S Eigenvectors : main directions Eigenvalue : variance along eigenvector
11
Eigenvector Decomposition If A is a square matrix then an eigenvector of A is a vector, p, such that Usually p is scaled to have unit length,|p|=1
12
Eigenvector Decomposition If K is an n x n covariance matrix, there exist n linearly independent eigenvectors, and all the corresponding eigenvalues are non- negative. We can decompose K as
13
Eigenvector Decomposition Recall that a normal pdf has The inverse of the covariance matrix is
14
Fun with Eigenvectors The normal distribution has form
15
Fun with Eigenvectors Consider the transformation
16
Fun with Eigenvectors The exponent of the distribution becomes
17
Normal distribution Thus by applying the transformation The normal distribution is simplified to
18
Dimensionality Reduction Co-ords often correllated Nearby points move together
19
Dimensionality Reduction Data lies in subspace of reduced dim. However, for some t,
20
Approximation Each element of the data can be written
21
Normal PDF
22
Useful Trick If x of high dimension, S huge If No. samples, N<dim(x) use
23
Building Eigen-Models Given examples Compute mean and eigenvectors of covar. Model is then P – First t eigenvectors of covar. matrix b – Shape model parameters
24
Eigen-Face models Model of variation in a region
25
Applications: Locating objects Scan window over target region At each position: Sample, normalise, evaluate p(g) Select position with largest p(g)
26
Multi-Resolution Search Train models at each level of pyramid Gaussian pyramid with step size 2 Use same points but different local models Start search at coarse resolution Refine at finer resolution
27
Application: Object Detection Scan image to find points with largest p(g) If p(g)>p min then object is present Strictly should use a background model: This only works if the PDFs are good approximations – often not the case
28
Back (sadly) to Texture Models raster scan Normalizations
29
PCA Galore Reduce Dimensions of shape vector Reduce Dimension of “texture” vector They are still correlated; repeat..
30
Object/Image to Parameters modeling ~80
31
Playing with the Parameters First two modes of shape variationFirst two modes of gray-level variation First four modes of appearance variation
32
Active Appearance Model Search Given: Full training model set, new image to be interpreted, “reasonable” starting approximation Goal: Find model with least approximation error High Dimensional Search: Curse of the dimensions strikes again
33
Active Appearance Model Search Trick: Each optimization is a similar problem, can be learnt Assumption: Linearity Perturb model parameters with known amount Generate perturbed image and sample error Learn multivariate regression for many such perterbuations
34
Active Appearance Model Search Algorithm: current estimate of model parameters: normalized image sample at current estimate
35
Active Appearance Model Search Slightly different modeling: Error term: Taylor expansion (with linear assumption) Min (RMS sense) error: Systematically perturb and estimate by numerical differentiation
36
Active Appearance Model Search (Results)
37
Sub-cortical Structures Initial PositionConverged
38
Random Aside Shape Vector provides alignment = 43 Alexei (Alyosha) Efros, 15-463 (15-862): Computational Photography, http://graphics.cs.cmu.edu/courses/15-463/2005_fall/www/Lectures/faces.ppt
39
Random Aside Alignment is the key 1. Warp to mean shape 2. Average pixels Alexei (Alyosha) Efros, 15-463 (15-862): Computational Photography, http://graphics.cs.cmu.edu/courses/15-463/2005_fall/www/Lectures/faces.ppt
40
Random Aside Enhancing Gender more same original androgynous more opposite D. Rowland, D. Perrett. “Manipulating Facial Appearance through Shape and Color”, IEEE Computer Graphics and Applications, Vol. 15, No. 5: September 1995, pp. 70-76
41
Random Aside (can’t escape structure!) Alexei (Alyosha) Efros, 15-463 (15-862): Computational Photography, http://graphics.cs.cmu.edu/courses/15-463/2005_fall/www/Lectures/faces.ppt Antonio Torralba & Aude Oliva (2002) Averages: Hundreds of images containing a person are averaged to reveal regularities in the intensity patterns across all the images.
42
Random Aside (can’t escape structure!) “100 Special Moments” by Jason Salavon Jason Salavon, http://salavon.com/PlayboyDecades/PlayboyDecades.shtml
43
Random Aside (can’t escape structure!) “Every Playboy Centerfold, The Decades (normalized)” by Jason Salavon 1960s1970s1980s Jason Salavon, http://salavon.com/PlayboyDecades/PlayboyDecades.shtml
Similar presentations
© 2024 SlidePlayer.com Inc.
All rights reserved.