de.mpg.biochem.mars.image

Class MarsImageUtils

java.lang.Object

de.mpg.biochem.mars.image.MarsImageUtils

public class MarsImageUtils
extends java.lang.Object

Common utility functions for DoG filtering, peak finding, peak fitting and related operations.

Constructor Summary

Constructors

Constructor and Description
MarsImageUtils()

Method Summary

Modifier and Type	Method and Description
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> double	calcR2(net.imglib2.RandomAccess<T> ra, int radius, Gaussian2D gauss) This method calculates the R-squared for the 2D gaussian provided within the radius specified for the image given as a RandomAccess.
static <T extends net.imglib2.type.numeric.RealType<T>> net.imglib2.RandomAccessibleInterval<net.imglib2.type.numeric.real.FloatType>	dogFilter(net.imglib2.RandomAccessibleInterval<T> img, double dogFilterRadius) This method uses the OpService to apply a Difference of Gaussian (DoG) filter on a 2D image provided.
static <T extends net.imglib2.type.numeric.RealType<T>> net.imglib2.RandomAccessibleInterval<net.imglib2.type.numeric.real.FloatType>	dogFilter(net.imglib2.RandomAccessibleInterval<T> img, double dogFilterRadius, int numThreads) This method uses the OpService to apply a Difference of Gaussian (DoG) filter on a 2D image provided.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak>	findPeaks(net.imglib2.Cursor<T> cursor, int t, double threshold, int minimumDistance, boolean findNegativePeaks) This method returned a list of peaks in the 2D image using the cursor specified that are above the pixel value threshold specified.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak>	findPeaks(net.imglib2.IterableInterval<T> iterableInterval, int t, double threshold, int minimumDistance, boolean findNegativePeaks) This method returned a list of peaks in the 2D image within the interval and iterable interval specified that are above the pixel value threshold specified.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak>	findPeaks(net.imglib2.RandomAccessible<T> img, net.imglib2.Interval interval, int t, double threshold, int minimumDistance, boolean findNegativePeaks) This method returned a list of peaks in the 2D image within the interval specified that are above the pixel value threshold specified.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak>	fitPeaks(net.imglib2.RandomAccessible<T> img, net.imglib2.Interval interval, java.util.List<Peak> peaks, int radius, double initialSize, boolean findNegativePeaks, double RsquaredMin) Given a 2D image and a list of peak positions in pixels, this function performs subpixel 2D gaussian fitting and updates the peak positions accordingly.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak>	fitPeaks(net.imglib2.RandomAccessible<T> img, net.imglib2.Interval interval, java.util.List<Peak> peaks, int radius, double initialSize, double fitRegionThreshold, boolean findNegativePeaks) Given a 2D image and a list of peak positions in pixels, this function performs subpixel 2D gaussian fitting and updates the peak positions accordingly.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> net.imglib2.RandomAccessibleInterval<T>	get2DHyperSlice(net.imagej.ImgPlus<T> img, int z, int c, int t) Convenience method to retrieve a 2D view from an ImgPlus for a given z, c, and t.
static java.util.List<int[]>	innerOffsets(int innerRadius) This method creates a list of integer x, y offsets from a central point based on the provided radius.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> void	integratePeaks(net.imglib2.RandomAccessible<T> img, net.imglib2.Interval interval, java.util.List<Peak> peaks, int innerRadius, int outerRadius) This method integrates the intensity of peaks in the 2D image provided.
static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> void	integratePeaks(net.imglib2.RandomAccessible<T> img, net.imglib2.Interval interval, java.util.List<Peak> peaks, int innerRadius, int outerRadius, boolean verbose) This method integrates the intensity of peaks in the 2D image provided.
static boolean	intervalContains(net.imglib2.Interval interval, double... location)
static java.util.List<int[]>	outerOffsets(int innerRadius, int outerRadius) This method creates a list of integer x, y offsets between the provided inner radius and outer radius.
static java.util.List<Peak>	removeNearestNeighbors(java.util.List<Peak> peaks, int minimumDistance) This method removes peaks from the provided list that are closer than the minimum distance given.
static net.imglib2.roi.IterableRegion<net.imglib2.type.logic.BoolType>	toIterableRegion(net.imglib2.roi.RealMask mask, net.imglib2.Interval image) This method converts from a RealMask to an IterableRegion
static net.imglib2.roi.IterableRegion<net.imglib2.type.logic.BoolType>	toIterableRegion(net.imglib2.roi.RealMaskRealInterval roi) This method converts from a RealMaskRealInterval to an IterableRegion

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

MarsImageUtils

public MarsImageUtils()

Method Detail

innerOffsets

public static java.util.List<int[]> innerOffsets(int innerRadius)

This method creates a list of integer x, y offsets from a central point based on the provided radius. The offset list can be used to efficiently visit the neighborhood of x, y positions around a central integration position.

Parameters:

innerRadius - Inner integration radius.

Returns:

The list of x, y offsets for innerRadius.

outerOffsets

public static java.util.List<int[]> outerOffsets(int innerRadius,
                                                 int outerRadius)

This method creates a list of integer x, y offsets between the provided inner radius and outer radius. The offset list can be used to efficiently visit the background neighborhood of x, y positions forming an outside ring around a central integration position.

Parameters:

innerRadius - Inner integration radius.

outerRadius - Outer integration radius.

Returns:

The list of x, y offsets for between innerRadius and outerRadius.

toIterableRegion

public static net.imglib2.roi.IterableRegion<net.imglib2.type.logic.BoolType> toIterableRegion(net.imglib2.roi.RealMaskRealInterval roi)

This method converts from a RealMaskRealInterval to an IterableRegion

Parameters:

roi - A RealMaskRealInterval represent the region of interest to convert.

Returns:

An IterableRegion.

toIterableRegion

public static net.imglib2.roi.IterableRegion<net.imglib2.type.logic.BoolType> toIterableRegion(net.imglib2.roi.RealMask mask,
                                                                                               net.imglib2.Interval image)

This method converts from a RealMask to an IterableRegion

Parameters:

mask - The RealMaks representing the roi.

image - The Interval of the image.

Returns:

An IterableRegion.

findPeaks

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak> findPeaks(net.imglib2.IterableInterval<T> iterableInterval,
                                                                                                                               int t,
                                                                                                                               double threshold,
                                                                                                                               int minimumDistance,
                                                                                                                               boolean findNegativePeaks)

This method returned a list of peaks in the 2D image within the interval and iterable interval specified that are above the pixel value threshold specified. The local maximum within the minimum distance is always chosen. The point in time provided is set for all peaks returned. Negative peaks can be located if desired. The peak search will only be performed in the iterable interval provided.

Type Parameters:

T - Image type.

Parameters:

iterableInterval - The IterableInterval to search for peaks in the image.

t - The T position being searched for peaks.

threshold - The pixel value threshold for peak detection.

minimumDistance - The minimum allowed distance between peaks.

findNegativePeaks - Whether to search for negative peaks.

Returns:

The list of peaks found.

findPeaks

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak> findPeaks(net.imglib2.RandomAccessible<T> img,
                                                                                                                               net.imglib2.Interval interval,
                                                                                                                               int t,
                                                                                                                               double threshold,
                                                                                                                               int minimumDistance,
                                                                                                                               boolean findNegativePeaks)

This method returned a list of peaks in the 2D image within the interval specified that are above the pixel value threshold specified. The local maximum within the minimum distance is always chosen. The point in time provided is set for all peaks returned. Negative peaks can be located if desired.

Type Parameters:

T - Image type.

Parameters:

img - 2D image containing peaks.

interval - The interval to search for peaks in the image.

t - The T position being searched for peaks.

threshold - The pixel value threshold for peak detection.

minimumDistance - The minimum allowed distance between peaks.

findNegativePeaks - Whether to search for negative peaks.

Returns:

The list of peaks found.

findPeaks

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak> findPeaks(net.imglib2.Cursor<T> cursor,
                                                                                                                               int t,
                                                                                                                               double threshold,
                                                                                                                               int minimumDistance,
                                                                                                                               boolean findNegativePeaks)

This method returned a list of peaks in the 2D image using the cursor specified that are above the pixel value threshold specified. The local maximum within the minimum distance is always chosen. The point in time provided is set for all peaks returned. Negative peaks can be located if desired.

Type Parameters:

T - Image type.

Parameters:

cursor - The cursor used to iterate through the regions of interest.

t - The T position being searched for peaks.

threshold - The pixel value threshold for peak detection.

minimumDistance - The minimum allowed distance between peaks.

findNegativePeaks - Whether to search for negative peaks.

Returns:

The list of peaks found.

dogFilter

public static <T extends net.imglib2.type.numeric.RealType<T>> net.imglib2.RandomAccessibleInterval<net.imglib2.type.numeric.real.FloatType> dogFilter(net.imglib2.RandomAccessibleInterval<T> img,
                                                                                                                                                       double dogFilterRadius)

This method uses the OpService to apply a Difference of Gaussian (DoG) filter on a 2D image provided. The dog filter op requires two sigmas that are used to generate two images and the difference of these images is returned. The relationship between dogFilterRadius and the two sigmas is the following: sigma1 = dogFilterRadius / sqrt(2) * 1.1 sigma2 = dogFilterRadius / sqrt(2) * 0.9

Type Parameters:

T - Image type.

Parameters:

img - 2D image that will be dog filtered.

dogFilterRadius - Radius to use for dog filtering.

Returns:

The dog filtered image.

dogFilter

public static <T extends net.imglib2.type.numeric.RealType<T>> net.imglib2.RandomAccessibleInterval<net.imglib2.type.numeric.real.FloatType> dogFilter(net.imglib2.RandomAccessibleInterval<T> img,
                                                                                                                                                       double dogFilterRadius,
                                                                                                                                                       int numThreads)

This method uses the OpService to apply a Difference of Gaussian (DoG) filter on a 2D image provided. The dog filter op requires two sigmas that are used to generate two images and the difference of these images is returned. The relationship between dogFilterRadius and the two sigmas is the following: sigma1 = dogFilterRadius / sqrt(2) * 1.1 sigma2 = dogFilterRadius / sqrt(2) * 0.9

Type Parameters:

T - Image type.

Parameters:

img - 2D image that will be dog filtered.

dogFilterRadius - Radius to use for dog filtering.

numThreads - The number of threads to use.

Returns:

The dog filtered image.

fitPeaks

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak> fitPeaks(net.imglib2.RandomAccessible<T> img,
                                                                                                                              net.imglib2.Interval interval,
                                                                                                                              java.util.List<Peak> peaks,
                                                                                                                              int radius,
                                                                                                                              double initialSize,
                                                                                                                              boolean findNegativePeaks,
                                                                                                                              double RsquaredMin)

Given a 2D image and a list of peak positions in pixels, this function performs subpixel 2D gaussian fitting and updates the peak positions accordingly. If the R-squared from fitting is below the threshold provided, the whole pixel position is left unchanged. The image is mirrored for pixel values beyond the interval provided. The starting guess for sigma is half initialSize.

Type Parameters:

T - Image type.

Parameters:

img - 2D image containing peaks.

interval - The interval to use for peak fitting.

peaks - The list of Peaks to fit with subpixel accuracy.

radius - The radius of the square region of pixels to use for fitting.

initialSize - A starting guess for the peak size.

findNegativePeaks - Whether negative peaks are being fit.

RsquaredMin - The minimum allowed R-squared value below which fits are rejected.

Returns:

The list of peaks after fitting with those having rejected fits removed.

fitPeaks

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> java.util.List<Peak> fitPeaks(net.imglib2.RandomAccessible<T> img,
                                                                                                                              net.imglib2.Interval interval,
                                                                                                                              java.util.List<Peak> peaks,
                                                                                                                              int radius,
                                                                                                                              double initialSize,
                                                                                                                              double fitRegionThreshold,
                                                                                                                              boolean findNegativePeaks)

Given a 2D image and a list of peak positions in pixels, this function performs subpixel 2D gaussian fitting and updates the peak positions accordingly. If the R-squared from fitting is below the threshold provided, the whole pixel position is left unchanged. The image is mirrored for pixel values beyond the interval provided. The starting guess for sigma is half initialSize.

Type Parameters:

T - Image type.

Parameters:

img - 2D image containing peaks.

interval - The interval to use for peak fitting.

peaks - The list of Peaks to fit with subpixel accuracy.

radius - The radius of the square region of pixels to use for fitting.

initialSize - A starting guess for the peak size.

fitRegionThreshold - The threshold pixel value for the region to fit.

findNegativePeaks - Whether negative peaks are being fit.

Returns:

The list of peaks after fitting with those having rejected fits removed.

calcR2

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> double calcR2(net.imglib2.RandomAccess<T> ra,
                                                                                                              int radius,
                                                                                                              Gaussian2D gauss)

This method calculates the R-squared for the 2D gaussian provided within the radius specified for the image given as a RandomAccess.

Type Parameters:

T - Image type.

Parameters:

ra - 2D image containing peak used for the R-squared calculation.

radius - The radius of the square region used for the R-squared calculation.

gauss - The 2D gaussian from fitting to use for the R-squared calculation.

Returns:

The R-squared value.

removeNearestNeighbors

public static java.util.List<Peak> removeNearestNeighbors(java.util.List<Peak> peaks,
                                                          int minimumDistance)

This method removes peaks from the provided list that are closer than the minimum distance given. If two peaks are closer than the minimum distance, the peak with the lower R-squared value, indicating a poorer fit, will be removed.

Parameters:

peaks - The list of peaks that nearest neighbors should be removed from.

minimumDistance - The smallest allowed distance between peaks in pixels.

Returns:

A new peak list with nearest neighbors removed.

integratePeaks

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> void integratePeaks(net.imglib2.RandomAccessible<T> img,
                                                                                                                    net.imglib2.Interval interval,
                                                                                                                    java.util.List<Peak> peaks,
                                                                                                                    int innerRadius,
                                                                                                                    int outerRadius)

This method integrates the intensity of peaks in the 2D image provided. The interval given is mirrored at the edges for pixel values that lie outside the boundaries. The values of all pixels within the region defined by the innerRadius are summed. The median value of the pixels between innerRadius and outerRadius is subtracted from each pixel in the sum to yield the background corrected total intensity.

Type Parameters:

T - Image type.

Parameters:

img - 2D image containing peaks.

interval - The interval to mirror at the edges during integration.

peaks - The Peaks to integrate.

innerRadius - The region to integrate.

outerRadius - The outer radius of the region used to calculate the background.

integratePeaks

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> void integratePeaks(net.imglib2.RandomAccessible<T> img,
                                                                                                                    net.imglib2.Interval interval,
                                                                                                                    java.util.List<Peak> peaks,
                                                                                                                    int innerRadius,
                                                                                                                    int outerRadius,
                                                                                                                    boolean verbose)

This method integrates the intensity of peaks in the 2D image provided. The interval given is mirrored at the edges for pixel values that lie outside the boundaries. The values of all pixels within the region defined by the innerRadius are summed. The median value of the pixels between innerRadius and outerRadius is subtracted from each pixel in the sum to yield the background corrected total intensity.

Type Parameters:

T - Image type.

Parameters:

img - 2D image containing peaks.

interval - The interval to mirror at the edges during integration.

peaks - The Peaks to integrate.

innerRadius - The region to integrate.

outerRadius - The outer radius of the region used to calculate the background.

verbose - Whether to include all values.

get2DHyperSlice

public static <T extends net.imglib2.type.numeric.RealType<T> & net.imglib2.type.NativeType<T>> net.imglib2.RandomAccessibleInterval<T> get2DHyperSlice(net.imagej.ImgPlus<T> img,
                                                                                                                                                        int z,
                                                                                                                                                        int c,
                                                                                                                                                        int t)

Convenience method to retrieve a 2D view from an ImgPlus for a given z, c, and t. Axis positions are based on a zero index. This method re-slices along c, t, and z. Therefore, for higher dimensional images, the output may be more than 2D.

Type Parameters:

T - Image type.

Parameters:

img - ImgPlus provide a view from.

z - The Z axis position.

c - The C axis position.

t - The T axis position.

Returns:

The slice of the image specified.

intervalContains

public static boolean intervalContains(net.imglib2.Interval interval,
                                       double... location)