Independent Component Analysis features of Color & Stereo
Download
Report
Transcript Independent Component Analysis features of Color & Stereo
Independent Component Analysis
features of Color & Stereo images
Authors:
Patrik O. Hoyer
Aapo Hyvarinen
CIS 526: Neural Computation
Presented by: Ajay Kumar Yadav
Overview
Introduction
Background Study
Data Preprocessing
Color Image Experiment
Stereo Image Experiment
Conclusion
Introduction
Visual Cortex: part of the cerebral cortex
responsible for processing visual stimuli.
(Static, Moving & Pattern Recognition)
Receptive fields are divided as:
Sub-regions that exert an excitatory influence.
(light grey)
Sub-regions that exert an inhibitory influence.
(dark grey)
Stimulus Influence also depends on size,
orientation and position
(Hubel & Wiesel’s -1962, DeValois-1982, DeAngelis-1993)
Contd..
Cones consist of three cell each responsible for each RGB
component. (tuned at wavelength of 430, 535, 590
nanometer)
The degree to which the images are non-corresponding
is defined as binocular disparity. It is used to determine
the distance of an object from oneself, and its relation to
the fixation plane, is called stereopsis.
Background Study
The sparseness-maximization network and
ICA are closely related. (Olshausen and
Field 1997)
n
x ai si
i 1
Hateren and Vander Schaaf qualitatively
compared the filter learned by ICA to
measurements of neural receptive fields.
Van Hateren and Ruderman proved ICA
also fit the receptive field properties for
video images.
Data Preprocessing
ICA preprocess the data in two steps:
The mean of the data is subtracted to center the
data on the origin.
x x E{x}
Whiten the data
T
z = Vx, so that E{zz } I
Goal: ICA transform W to minimize the statistical
dependencies between the estimated sources.
sˆ Wz WVPCA x WDn1/ 2 EnT x
After convergence
A En Dn1/ 2W T
Color Image Experiment
Standard RGB values are considered as input
data assuming the transformation to cone
outputs to be roughly linear.
A total of 50,000 12 by 12 pixel image patches
were sampled randomly with dimensionality of
432.
Data is preprocessed and correlation matrix
and eigen vectors are calculated.
Constant RGB value is used in the display.
Correlation matrix
Data is projected in 160
principle component
before whitening. Two
reasons are:
To emulate the real
neuron functionality
Dimension is dropped to
lower computational
cost.
Results
**ICA basis of color images**
**Color content of three ICA filters**
**Percentage of achromatic**
Stereo Image Experiment
Stereo image data:
5 focus points at random from each image are
selected and estimated the disparities.
Randomly sampled 16*16 pixel corresponding
image in patch area of 300*300 pixels centered
on each focus point.
Due to the fluctuation patches are often similar
but horizontally shifted.
During the preprocessing local mean was
removed from each component and correlation
matrix and eigenvalue decomposition are
calculated.
Stereo Images
**PCA Basis of Stereo Image**
Equal Response
Varying Response
**ICA Basis of Stereo Image**
Ocular Dominance
The shift from one eye to the other takes place over a
distance of less than 50 microns, therefore column
dominated by one eye.
If the sampling areas is smaller, correlation between the
patches would be higher.
If the area gets larger, the dependencies between the
left and right patches get weaker
Disparity Tuning Analysis
To analyze the
disparity tuning
several ICA basis were
estimated using
different number
random seeds.
Only relatively high
frequency well
localized binocular
vectors are selected
Disparity Tuning Curves
Each patch is shown to both
the eyes to get the tuning
curve and the mean is
considered as final curve.
These curves are defined in
two parts:
Tuned excitatory
Tuned inhibitory
Tuned excitatory shows a
strong peak at zero.
Tuned inhibitory shows
opposite polarity.
Near unit’s right receptive
slightly shifted giving positive
preferred disparity.
Far unit has opposite
positional offset with negative
disparities.
Conclusion
ICA could be applied in denoising,
compression or pattern recognition of color
or stereo data.
ICA can be used to model computational
properties of visual cortex (V1) cell.
Limitation:
Since ICA emulate the behavior of cones it may
fail in dark or un-illuminated images.
To get better correlation basis patch needed to
be small which may vary.
Nonlinearities inherent in the conversion from
RGB to cones response will affect the ICA result.