New dictionary and fast atom searching method for matching pursuit representation of displaced frame difference
read more
Citations
Denoising by sparse approximation: error bounds based on rate-distortion theory
Matching pursuits coding of data
Video encoding and decoding methods and corresponding devices
Basis picking for matching pursuits image coding
Low complexity separable matching pursuits
References
Matching pursuits with time-frequency dictionaries
Very low bit-rate video coding based on matching pursuits
Video compression using matching pursuits
Efficient image representation by anisotropic refinement in matching pursuit
Redundancy in non-orthogonal transforms
Related Papers (5)
Frequently Asked Questions (13)
Q2. What is the main limitation in the adoption of a redundant dictionary?
Matching Pursuit (MP) algorithms iteratively decompose a signal in its most important features using a set of atoms chosen among a redundant dictionary of basis functions.
Q3. What is the effect of anisotropy on the dictionary?
Anisotropy increases the redundancy in the dictionary because of the introduction of an extra parameter to code, but as it has been shown in [3] this produces an overall increase in efficiency.
Q4. how many atoms are in the dictionary?
2/1221 1 βα . (5) α is an optimality factor related to the strategy adopted to determine the best atom in the dictionary, while β strictly depends on the dictionary representing its ability to capture the feature of the input function f [5].
Q5. What is the assumption that the authors made?
The assumption the authors made is that the value of the scalar product between the residual image and the atom increases as soon as the authors are getting closer to the right angle, for most of the atoms.
Q6. how many steps can be used to decompose a dictionary?
Applying iterativelysuch a procedure, after N iterations the authors obtain:fRgfRgf N Nnn nn += ∑ −=1 0 , γγ (3)where fRn is the residual at the nth step and ffR =0 .
Q7. What are the advantages of non orthogonal transforms?
Non orthogonal transforms represent indeed a valid alternative to orthogonal transforms like DCT or wavelet based scheme especially at low bit-rates, where most of the signal energy can be captured by only few elements.
Q8. What is the simplest way to build a dictionary?
This dictionary is built acting on a generating function of unit L2 norm by means of a family of unitary operators Uγ:},{ Γ∈= γγUD , (6) for a given set of indexes Γ .
Q9. How can the authors improve the coding efficiency of video?
In [1, 2] authors have shown that improved coding efficiency can be achieved by replacing the DCT with an overcomplete transform.
Q10. What is the angle to search?
Once found the best matching angle, then a dichotomist process starts which keeps on dividing by two the angle until the authors get to the unit angle γ=Π/128.
Q11. What is the gain G in terms of complexity of the proposed approach when compared to the full?
Indicating with N=128 the possible angles to be searched, then the complexity Nr of the proposed search can be expressed as follows: Nr = 4 + 2log2 δ/ γ . (11) The gain G in terms of complexity of the proposed approach when compared to the full search is the following.
Q12. How much would it cost to code another dictionary?
the cost of coding another dictionary would be just one bit per atom, while adaptive dictionaries could probably better match the evolving structure of the residual.
Q13. Why does the g0 algorithm show a better MSE?
Probably this is due to the fact that the peaky generating function g0 suits particularly well when the atom is set in the position where there is a peak in the residual.