Frequency Flashcards
Basis for vector space of dimension n
N vectors are independent
N vectors span the vector space
Orthonormal basis vectors
All bias vectors are if length 1
All pairs of basis vectors are orthogonal
Expansion is only down by projection
If a neighborhood N has large dot product with a basis vector (image), then
N is similar to that basis vector image
Feel-Chen basis
Vectors to represent
Gradient, ripple, line, Laplacian, and constant features
Fourier series
Any univariate function can be rewritten as a weighted sum of wines and cosines of different frequencies
Discrete Fourier transform
NxN spatial samples to F(u,v) array of coefficients for frequency representation
Inverse DFT
Frequency representation to spatial samples
Why do we use Fourier basis
Removes high frequency noise
Extract texture features
Image compression