registration
Class RegisterRetina
java.lang.Object
|
+--registration.Register
|
+--registration.RegisterMI
|
+--registration.RegisterRetina
- public class RegisterRetina
- extends RegisterMI
Retina registration methods.
(c) 2003 Michael Abramoff. All rights reserved.
Constructor Summary |
RegisterRetina(float[] reference,
int width,
float[] mask)
Prepare for registration including preprocessing of the reference image. |
Method Summary |
static float[] |
featureBlobs(float[] image,
int width)
|
static float[] |
featureEdges(float[] image,
int width)
|
static void |
makeMaskedNaN(float[] image,
float[] mask)
Mask the pixels in a vector image by setting all pixels that have mask == 0 to NaN. |
float[] |
register(float[] a)
Register the image a to the reference image. |
Methods inherited from class java.lang.Object |
, clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait |
referenceDisk
protected float[] referenceDisk
RegisterRetina
public RegisterRetina(float[] reference,
int width,
float[] mask)
- Prepare for registration including preprocessing of the reference image.
The reference image will be used in register.
- Parameters:
image
- an imagewidth
- the width of the imagemask
- a float[] of the same size as reference and a containing a 1 for each valid pixel, 0 if not valid.
register
public float[] register(float[] a)
- Register the image a to the reference image. First make an estimate of the disk using blobs, then
refine the estimate using MI and Powell's minimization.
Uses detection of the disk to minimize local minima in the minimization and false registrations.
If mask is given, use only the pixels in a that are 1 in mask, and disregard the pixels that are 0 in mask.
- Overrides:
register
in class RegisterMI
- Parameters:
a
- an image- Returns:
- a vector with the registration coefficients
featureBlobs
public static float[] featureBlobs(float[] image,
int width)
featureEdges
public static float[] featureEdges(float[] image,
int width)
makeMaskedNaN
public static void makeMaskedNaN(float[] image,
float[] mask)
- Mask the pixels in a vector image by setting all pixels that have mask == 0 to NaN.