endrov.unsortedImageFilters
Class GeometricMomentum
java.lang.Object
endrov.unsortedImageFilters.GeometricMomentum
public class GeometricMomentum
- extends java.lang.Object
Geometric momentum, defined as m_pq=sum_ij x^p y^q f(x,y), can be used for shape recognition and localization.
Central moment is calculated as: mu_pq=sum_ij (x-x_0)^p (y-y_0)^q f(x,y) with (x_0,y_0)=(m_10,m_01)/m_00, that is, the geometric center.
See "Moment functions in image analysis", R Makundan, K R Ramakrishnan
Examples:
m_pq/(m_00^((p+q+2)/2)) is scale invariant
mu_pq/(mu_00^((p+q+2)/2)) is scale and translation invariant
rotation invariants exists.
Method Summary |
static double |
momentum(ProgressHandle ph,
EvPixels in,
int p,
int q)
Calculate momentum: m_pq=sum_ij x^p y^q f(x,y)
TODO special faster versions can be deployed if image is binary
Complexity O(w*h*(p+q)) |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
GeometricMomentum
public GeometricMomentum()
momentum
public static double momentum(ProgressHandle ph,
EvPixels in,
int p,
int q)
- Calculate momentum: m_pq=sum_ij x^p y^q f(x,y)
TODO special faster versions can be deployed if image is binary
Complexity O(w*h*(p+q))