A reweighting class where the objective function tends to a fixed value, rather than infinity. More...
#include <irls.h>
Collaboration diagram for TooN::RobustIII< Precision >:
Public Member Functions | |
| void | set_sd (Precision x) |
| Set the noise standard deviation. | |
| Precision | reweight (Precision x) const |
| Returns \(w(x)\). More... | |
| Precision | objective (Precision x) const |
Public Attributes | |
| Precision | sd_inlier |
| Inlier standard deviation squared. More... | |
Detailed Description
template<typename Precision>
struct TooN::RobustIII< Precision >
A reweighting class where the objective function tends to a fixed value, rather than infinity.
Note that this is not therefore a proper distribution since its integral is not finite. It is considerably more efficient than RobustI and II, since log() is not used.
Member Function Documentation
◆ reweight()
template<typename Precision >
|
inline |
Returns \(w(x)\).
Returns \(\int xw(x)dx\).
Member Data Documentation
◆ sd_inlier
template<typename Precision >
| Precision TooN::RobustIII< Precision >::sd_inlier |
Inlier standard deviation squared.
The documentation for this struct was generated from the following file:
- TooN/irls.h
