some matlab scripts:
builds an image from the rgb component matrices
takes 3 component matrices and builds one 3XN matrix
that represents a multivariate distribution (matrix of observations).
opposite of above
Takes a vector representing a discrete probability density function (PDF)
and returns the cumulative density function (CDF)
Takes a vector and builds a circulent matrix using this vector as the
Takes a matrix representing the joint probability distribution of two
discrete random variables X(rows) Y(cols) and returns conditional
information Hx|y and Hy|x
Takes an integer argument (N) and returns a Fourier transform matrix of
Takes an integer argument (N) and returns a low pass filter of length N
Takes an integer argument (N) and returns a high pass filter of length N
Takes an integer argument (N) and a vector representing a PDF, returns
a vextor of length N representing the value of the inverse CDF at N equally
spaced points between 0 and 1.
takes a pair of images of equal size and returns the 2d joint histogram.
Takes a matrix and its covariance matrix and returns the KL transform
takes a matrix representing a joint histogram returns the marginal
histograms Hx and Hy
Takes a pair of images and returns the mutual information Ixy
Normalizes a matrix (returns values between 0 and 1)
Takes a vector p representing a 1D PDF and an integer N, returns a NxN
matrix of random numbers with the distribution p
Downsamples a vector.
Upsamples a vector inserting zeros.
A script that simulates two channel subband codling based on the "ideal"
lower (H0) and upper (H1) filters.
subband coding script
A script takes the rgb components of an image and treats them as samples
of a 3x1 vector random variable, computes the zero mean image, computes
the covariance matrix, uses the eigenvectors to decorrelate rgb ( the
KL transform ) rebuilds and displays new image.