All Packages  Class Hierarchy  This Package  Previous  Next  Index

Class rsfF.solver.MissingDataEquations 

java.lang.Object
   |
   +----rsfF.solver.MissingDataEquations
public class MissingDataEquations
extends Object
MissingDataEquations sets up the missing data problem:
A(u + k) = 0
That is, given known data, k, we fill in values of u which makes A(u + k) as close to zero as possible.
So as not to get the trivial solution u = -k, we restrict the search space of u using a masking operator. By using the default masking operator, the search space is the space in which the non-zero components of k are clamped to zero.

The actual form of the equation is:
AMu = -Ak
M is the free-data masking operator, k is the known data, and u is the known data for which we solve.

Use getAM() and getNegAk() to retrieve the operator AM and the Vector -Ak.
It is important to make sure that initially u = Mu, so if there is any doubt, use zeroKnownComponents(u).

Constructor Index

o MissingDataEquations(Operator, Rsf)
Assumes that the unknown data occurs whereever knownData is zero.
o MissingDataEquations(Operator, Vector, Mask)
This constructor allows known values to be zero by letting you specify the "freeMask", which zeroes all the components of an Rsf which correspond to known data components.

Method Index

o getAM()
o getNegAk()
o zeroKnownComponents(Vector)
sets u = Mu (where M zeroes the known data components)

Constructors

o MissingDataEquations 
 public MissingDataEquations(Operator A,
                             Rsf knownData)
Assumes that the unknown data occurs whereever knownData is zero.
If you need to specify that the data is known to be zero at certain places, use the constructor which specifies your own mask.

o MissingDataEquations 
 public MissingDataEquations(Operator A,
                             Vector knownData,
                             Mask freeMask)
This constructor allows known values to be zero by letting you specify the "freeMask", which zeroes all the components of an Rsf which correspond to known data components.

Methods

o getAM 
 public CompoundLinearOperatorAdjoint getAM()
Returns:
A(M) (where M is the operator which zeroes the known data components)
o getNegAk 
 public Vector getNegAk()
Returns:
-Ak (where k is the known data)
o zeroKnownComponents 
 public void zeroKnownComponents(Vector u)
sets u = Mu (where M zeroes the known data components) 

All Packages  Class Hierarchy  This Package  Previous  Next  Index