#include <LossFunctions.h>
Public Types | |
| typedef std::shared_ptr< L2WithDeadZone > | shared_ptr |
 Public Types inherited from gtsam::noiseModel::mEstimator::Base | |
| enum | ReweightScheme { Scalar, Block } |
| typedef std::shared_ptr< Base > | shared_ptr |
Public Member Functions | |
| bool | equals (const Base &expected, double tol=1e-8) const override |
| L2WithDeadZone (double k=1.0, const ReweightScheme reweight=Block) | |
| double | loss (double distance) const override |
| double | modelParameter () const |
| void | print (const std::string &s) const override |
| double | weight (double distance) const override |
| Public Member Functions inherited from gtsam::noiseModel::mEstimator::Base | |
| Base (const ReweightScheme reweight=Block) | |
| void | reweight (Vector &error) const |
| void | reweight (std::vector< Matrix > &A, Vector &error) const |
| void | reweight (Matrix &A, Vector &error) const |
| void | reweight (Matrix &A1, Matrix &A2, Vector &error) const |
| void | reweight (Matrix &A1, Matrix &A2, Matrix &A3, Vector &error) const |
| ReweightScheme | reweightScheme () const |
| Returns the reweight scheme, as explained in ReweightScheme. More... | |
| double | sqrtWeight (double distance) const |
| Vector | sqrtWeight (const Vector &error) const |
| Vector | weight (const Vector &error) const |
| virtual | ~Base () |
Static Public Member Functions | |
| static shared_ptr | Create (double k, const ReweightScheme reweight=Block) |
Protected Attributes | |
| double | k_ |
| Protected Attributes inherited from gtsam::noiseModel::mEstimator::Base | |
| ReweightScheme | reweight_ |
| Strategy for reweighting. More... | |
Detailed Description
L2WithDeadZone implements a standard L2 penalty, but with a dead zone of width 2*k, centered at the origin. The resulting penalty within the dead zone is always zero, and grows quadratically outside the dead zone. In this sense, the L2WithDeadZone penalty is "robust to inliers", rather than being robust to outliers. This penalty can be used to create barrier functions in a general way.
This model has a scalar parameter "k".
- Loss (x) = 0 if |x|<k, 0.5(k-|x|)² otherwise
- Derivative (x) = 0 if |x|<k, (-k+x) if x>k, (k+x) if x<-k
- Weight w(x) = (x)/x = 0 if |x|<k, (-k+x)/x if x>k, (k+x)/x if x<-k
Definition at line 446 of file LossFunctions.h.
Member Typedef Documentation
◆ shared_ptr
| typedef std::shared_ptr<L2WithDeadZone> gtsam::noiseModel::mEstimator::L2WithDeadZone::shared_ptr |
Definition at line 451 of file LossFunctions.h.
Constructor & Destructor Documentation
◆ L2WithDeadZone()
| gtsam::noiseModel::mEstimator::L2WithDeadZone::L2WithDeadZone | ( | double | k = 1.0, |
| const ReweightScheme | reweight = Block |
||
| ) |
Definition at line 392 of file LossFunctions.cpp.
Member Function Documentation
◆ Create()
|
static |
Definition at line 423 of file LossFunctions.cpp.
◆ equals()
|
overridevirtual |
Implements gtsam::noiseModel::mEstimator::Base.
Definition at line 417 of file LossFunctions.cpp.
◆ loss()
|
overridevirtual |
This method is responsible for returning the total penalty for a given amount of error. For example, this method is responsible for implementing the quadratic function for an L2 penalty, the absolute value function for an L1 penalty, etc.
TODO(mikebosse): When the loss function has as input the norm of the error vector, then it prevents implementations of asymmeric loss functions. It would be better for this function to accept the vector and internally call the norm if necessary.
This returns (x) in mEstimator
Reimplemented from gtsam::noiseModel::mEstimator::Base.
Definition at line 408 of file LossFunctions.cpp.
◆ modelParameter()
|
inline |
Definition at line 459 of file LossFunctions.h.
◆ print()
|
overridevirtual |
Implements gtsam::noiseModel::mEstimator::Base.
Definition at line 413 of file LossFunctions.cpp.
◆ weight()
|
overridevirtual |
This method is responsible for returning the weight function for a given amount of error. The weight function is related to the analytic derivative of the loss function. See https://members.loria.fr/MOBerger/Enseignement/Master2/Documents/ZhangIVC-97-01.pdf for details. This method is required when optimizing cost functions with robust penalties using iteratively re-weighted least squares.
This returns w(x) in mEstimator
Implements gtsam::noiseModel::mEstimator::Base.
Definition at line 399 of file LossFunctions.cpp.
Member Data Documentation
◆ k_
|
protected |
Definition at line 448 of file LossFunctions.h.
The documentation for this class was generated from the following files: