Public Member Functions | Protected Attributes

angle_state< observation_type > Class Template Reference

#include <angle_state.h>

Inheritance diagram for angle_state< observation_type >:
skeleton_state< observation_type > state< observation_type > brownian_state< observation_type > brownian_stick_state< observation_type > KBK_stick_state< observation_type > pik_state< observation_type > pik_stick_state< observation_type > tangent_state< observation_type >

List of all members.

Public Member Functions

 angle_state (const options &opts, const calibration &_calib, const bool _for_show=true)
double predict (const observation_type &observation, const double variance_scale=1.0)
angle_state & operator= (angle_state &new_state)
bool load (const std::string filename)
bool load (std::ifstream &file)

Protected Attributes

solver_type::vector_type prev_theta
const double pred_noise
const von_mises noise
const bool extrapolate

Detailed Description

template<class observation_type>
class angle_state< observation_type >

An implementation of the standard Gaussian distributed joint angles prior

\[ p(\theta_t | \theta_{t-1}) = \mathcal{N} (\theta_t | \theta_{t-1}, \Sigma) \]

where $\theta$ denotes the joint angles, subscript denotes temporal index and $\Sigma$ is a diagonal covariance. If you use this class we recommend citing 

 @inproceedings{Sidenbladh00a,
    Author = {Hedvig Sidenbladh and Michael J. Black and David J. Fleet},
    Booktitle = {ECCV},
    Pages = {702--718},
    Publisher = {Springer},
    Series = {LNCS 1843},
    Title = {Stochastic tracking of 3D human figures using 2D image motion},
    Volume = {II},
    Year = {2000}}

Constructor & Destructor Documentation

template<class observation_type>
angle_state< observation_type >::angle_state ( const options &  opts,
const calibration &  _calib,
const bool  _for_show = true 
) [inline]

Member Function Documentation

template<class observation_type>
bool angle_state< observation_type >::load ( const std::string  filename  )  [inline]

Load pose parameters from a file.

Parameters:
[in] filename The name of the file containing the pose paramters.
Returns:
TRUE on succes.

Reimplemented from skeleton_state< observation_type >.

Reimplemented in brownian_state< observation_type >, brownian_stick_state< observation_type >, pik_state< observation_type >, pik_stick_state< observation_type >, and tangent_state< observation_type >.

template<class observation_type>
bool angle_state< observation_type >::load ( std::ifstream &  file  )  [inline]

Load pose parameters from a stream. This is useful for batch loads.

Parameters:
[in] file A stream representing the open file.
Returns:
TRUE on succes.

Reimplemented from skeleton_state< observation_type >.

Reimplemented in brownian_state< observation_type >, brownian_stick_state< observation_type >, pik_state< observation_type >, pik_stick_state< observation_type >, and tangent_state< observation_type >.

template<class observation_type>
angle_state& angle_state< observation_type >::operator= ( angle_state< observation_type > &  new_state  )  [inline]
template<class observation_type>
double angle_state< observation_type >::predict ( const observation_type &  observation,
const double  variance_scale = 1.0 
) [inline, virtual]

The importance distribution of the tracker.

Parameters:
[in] observation The current observation.
[in] variance_scale A scale factor for the default variance. This is used for annealed filters.
Returns:
The correction factor for the importance distribution (most often 1).

Implements state< observation_type >.

Reimplemented in brownian_state< observation_type >, brownian_stick_state< observation_type >, KBK_stick_state< observation_type >, pik_state< observation_type >, pik_stick_state< observation_type >, tangent_state< observation_type >, and KBK_stick_state< point_cloud_and_stick >.

Member Data Documentation

template<class observation_type>
const bool angle_state< observation_type >::extrapolate [protected]
template<class observation_type>
const von_mises angle_state< observation_type >::noise [protected]
template<class observation_type>
const double angle_state< observation_type >::pred_noise [protected]
template<class observation_type>
solver_type::vector_type angle_state< observation_type >::prev_theta [protected]
The documentation for this class was generated from the following file:
Generated on Thu Dec 1 2011 12:54:08 for HUMIM Tracker by  doxygen 1.7.1