IVALab Python Libraries
Collection of code for computer vision and robotics with specific API.
Classes
Track point algorithms.

A collection of point tracking implementations for images. More...

Classes

class  centroid
 Used to track the centroid of a target object. More...
 
class  CfgCentMulti
 Configuration setting specifier for centroidMulti. More...
 
class  CfgCentroid
 Configuration setting specifier for centroidMulti. More...
 
class  TrackState
 Track state structure. More...
 

Detailed Description

A collection of point tracking implementations for images.

Encapsulates the generic functionality of a track pointer through this package, typically some type of segmentation-based track pointer. Other detecton schemes typically output track state or sub-state information. If sub-state information is provided, then usually tacking on a filter as part of a Perceiver is the way to go. The filter block can also be a smoother, just note that the best estimate is going to be a delayed one. Using a filter has the same properties as filter output has gain and phase properties that usually mean the signal is attentuated and delayed relative to the truth. Incorporating a good motion model/prior reduces the negative gain/phase properties, but usually that information is not known.