@Override
@SuppressWarnings("unchecked")
public O calculate(final I kernel, final Dimensions paddedDimensions) {
Dimensions paddedFFTInputDimensions;
// if an fftsize op has been set recompute padded size
if (fftSizeOp != null) {
long[][] sizes = fftSizeOp.calculate(paddedDimensions);
paddedFFTInputDimensions = new FinalDimensions(sizes[0]);
} else {
paddedFFTInputDimensions = paddedDimensions;
}
// compute where to place the final Interval for the kernel so that the
// coordinate in the center
// of the kernel is shifted to position (0,0).
final Interval kernelConvolutionInterval =
paddingIntervalCentered.calculate(kernel, paddedFFTInputDimensions);
final Interval kernelConvolutionIntervalOrigin =
paddingIntervalOrigin.calculate(kernel, kernelConvolutionInterval);
return (O)
Views.interval(
Views.extendPeriodic(
Views.interval(
Views.extendValue(kernel, Util.getTypeFromInterval(kernel).createVariable()),
kernelConvolutionInterval)),
kernelConvolutionIntervalOrigin);
}
@Test
public void testRandomAccessibleView() {
@SuppressWarnings("unchecked")
final RandomAccessible<ByteType> res =
(RandomAccessible<ByteType>)
ops.run(MapViewRandomAccessToRandomAccess.class, in, op, new ByteType());
final Cursor<ByteType> iterable = Views.iterable(Views.interval(res, in)).cursor();
while (iterable.hasNext()) {
assertEquals((byte) 10, iterable.next().get());
}
}
public static <T extends Type<T> & Comparable<T>> int countLocalMaxima(
final RandomAccessibleInterval<T> img, final Shape shape) {
int nMaxima = 0;
final RandomAccessibleInterval<T> source = Views.interval(img, Intervals.expand(img, -1));
final Cursor<T> center = Views.iterable(source).cursor();
A:
for (final Neighborhood<T> neighborhood : shape.neighborhoods(source)) {
final T c = center.next();
for (final T t : neighborhood) if (t.compareTo(c) > 0) continue A;
++nMaxima;
}
return nMaxima;
}
@SuppressWarnings("unchecked")
@Override
public boolean process() {
final long start = System.currentTimeMillis();
final ArrayImgFactory<FloatType> factory = new ArrayImgFactory<FloatType>();
final ArrayImg<FloatType, FloatArray> floatImage =
(ArrayImg<FloatType, FloatArray>) factory.create(source, new FloatType());
// Copy to float.
final Cursor<FloatType> cursor = floatImage.cursor(source);
final RandomAccess<T> ra = source.randomAccess(source);
while (cursor.hasNext()) {
cursor.fwd();
ra.setPosition(cursor);
cursor.get().set(ra.get().getRealFloat());
}
// Create result holders.
Dx = (ArrayImg<FloatType, FloatArray>) factory.create(source, new FloatType());
Dy = (ArrayImg<FloatType, FloatArray>) factory.create(source, new FloatType());
final int ndims = floatImage.numDimensions();
if (ndims == 3) {
final long nz = floatImage.dimension(2);
for (int z = 0; z < nz; z++) {
final IntervalView<FloatType> slice = Views.hyperSlice(floatImage, 2, z);
final IntervalView<FloatType> targetSliceX = Views.hyperSlice(Dx, 2, z);
final IntervalView<FloatType> targetSliceY = Views.hyperSlice(Dy, 2, z);
final boolean ok = processSlice(slice, targetSliceX, targetSliceY);
if (!ok) {
return false;
}
}
} else {
final boolean ok = processSlice(floatImage, Dx, Dy);
if (!ok) {
return false;
}
}
components.clear();
components.add(Dx);
components.add(Dy);
final long end = System.currentTimeMillis();
processingTime = end - start;
return true;
}
/**
* Computes a Gaussian convolution with double precision on an entire {@link Img}
*
* @param sigma - the sigma for the convolution
* @param img - the img {@link Img}
* @param outofbounds - the {@link OutOfBoundsFactory}
* @return the convolved img in {@link DoubleType}
*/
public static <T extends RealType<T>> Img<DoubleType> toDouble(
final double[] sigma,
final Img<T> img,
final OutOfBoundsFactory<DoubleType, RandomAccessibleInterval<DoubleType>> outofbounds) {
GaussDouble gauss = null;
try {
if (DoubleType.class.isInstance(img.firstElement())) {
@SuppressWarnings({"rawtypes", "unchecked"})
final Img<DoubleType> img2 = (Img) img;
gauss = new GaussDouble(sigma, img2);
} else {
final RandomAccessibleInterval<DoubleType> rIn =
new WriteConvertedIterableRandomAccessibleInterval<T, DoubleType, Img<T>>(
img, new RealDoubleSamplerConverter<T>());
gauss =
new GaussDouble(
sigma,
Views.extend(rIn, outofbounds),
img,
img.factory().imgFactory(new DoubleType()));
}
} catch (final IncompatibleTypeException e) {
return null;
}
gauss.call();
return (Img<DoubleType>) gauss.getResult();
}
/**
* Computes a Gaussian convolution with double precision on an entire {@link Img}
*
* @param sigma - the sigma for the convolution
* @param img - the img {@link Img}
* @param outofbounds - the {@link OutOfBoundsFactory}
* @return the convolved img having the input type
*/
@SuppressWarnings({"unchecked", "rawtypes"})
public static <T extends RealType<T>> Img<T> inDouble(
final double[] sigma,
final Img<T> img,
final OutOfBoundsFactory<DoubleType, RandomAccessibleInterval<DoubleType>> outofbounds) {
try {
if (DoubleType.class.isInstance(img.firstElement())) {
return (Img) toDouble(sigma, img, outofbounds);
}
final Img<T> output = img.factory().create(img, img.firstElement());
final RandomAccessible<DoubleType> rIn =
Views.extend(
new WriteConvertedRandomAccessibleInterval<T, DoubleType>(
img, new RealDoubleSamplerConverter<T>()),
outofbounds);
final RandomAccessible<DoubleType> rOut =
new WriteConvertedRandomAccessible<T, DoubleType>(
output, new RealDoubleSamplerConverter<T>());
inDouble(
sigma,
rIn,
img,
rOut,
new Point(sigma.length),
img.factory().imgFactory(new DoubleType()));
return output;
} catch (final IncompatibleTypeException e) {
return null;
}
}
/**
* Computes a Gaussian convolution in-place (temporary imgs are necessary) with the precision of
* the type provided on an entire {@link Img}
*
* @param sigma - the sigma for the convolution
* @param img - the img {@link Img} that will be convolved in place
*/
public static <T extends NumericType<T>> void inNumericTypeInPlace(
final double[] sigma,
final Img<T> img,
final OutOfBoundsFactory<T, RandomAccessibleInterval<T>> outofbounds) {
inNumericType(
sigma, Views.extend(img, outofbounds), img, img, new Point(sigma.length), img.factory());
}
@Override
protected IterableIntervalProjector2D<T, ARGBType> getProjector(
final int dimX,
final int dimY,
final RandomAccessibleInterval<T> source,
final RandomAccessibleInterval<ARGBType> target) {
return new IterableIntervalProjector2D<T, ARGBType>(
dimX, dimY, source, Views.iterable(target), m_converter);
}
/**
* Walk through an Interval on an Img using a Cursor
*
* @param img
* @param interval
*/
protected static void walkThrough(final Img<IntType> img, final Interval interval) {
final Cursor<IntType> c = Views.interval(img, interval).cursor();
int i = 0;
while (c.hasNext()) {
c.fwd();
i += c.get().get();
}
j = i;
}
/**
* Computes a Gaussian convolution with the precision of the type provided on an entire {@link
* Img}
*
* @param sigma - the sigma for the convolution
* @param img - the img {@link Img}
* @param outofbounds - the {@link OutOfBoundsFactory}
* @return the convolved img
*/
public static <T extends NumericType<T>> Img<T> inNumericType(
final double[] sigma,
final Img<T> img,
final OutOfBoundsFactory<T, RandomAccessibleInterval<T>> outofbounds) {
final Img<T> output = img.factory().create(img, img.firstElement());
inNumericType(
sigma, Views.extend(img, outofbounds), img, output, new Point(sigma.length), img.factory());
return output;
}
public FirstIteration(
final ImagePortion portion,
final RandomAccessibleInterval<FloatType> psi,
final ArrayList<RandomAccessibleInterval<FloatType>> imgs,
final ArrayList<RandomAccessibleInterval<FloatType>> weights) {
this.portion = portion;
this.psi = psi;
this.imgs = imgs;
this.weights = weights;
this.psiIterable = Views.iterable(psi);
this.iterableImgs = new ArrayList<IterableInterval<FloatType>>();
this.iterableWeights = new ArrayList<IterableInterval<FloatType>>();
this.realSum = new RealSum();
if (imgs.size() != weights.size())
throw new RuntimeException("Number of weights and images must be equal.");
compatibleIteration = true;
for (final RandomAccessibleInterval<FloatType> img : imgs) {
final IterableInterval<FloatType> imgIterable = Views.iterable(img);
if (!psiIterable.iterationOrder().equals(imgIterable.iterationOrder()))
compatibleIteration = false;
this.iterableImgs.add(imgIterable);
}
for (final RandomAccessibleInterval<FloatType> weight : weights) {
final IterableInterval<FloatType> weightIterable = Views.iterable(weight);
if (!psiIterable.iterationOrder().equals(weightIterable.iterationOrder()))
compatibleIteration = false;
for (final IterableInterval<FloatType> imgIterable : iterableImgs)
if (!imgIterable.iterationOrder().equals(weightIterable.iterationOrder()))
compatibleIteration = false;
this.iterableWeights.add(weightIterable);
}
}
/**
* Compute the steepest descent images of the template at the identity warp. Each steepest descent
* image comprises the partial derivatives of template intensities with respect to one parameter
* of the warp function.
*
* <p>The result is stored in the <em>n+1</em> dimensional {@link #target} image. Dimension
* <em>n</em> is used to index the partial derivative. For example, the partial derivative by the
* second parameter of the warp function is stored in slice <em>n=1</em>.
*
* @param gradients n+1 dimensional image of partial derivatives of the template. Dimension n is
* used to index the partial derivative. For example, the partial derivative by Y is stored in
* slice n=1.
* @param warpFunction The warp function to be applied to the template. The partial derivatives of
* template intensities with respect to the parameters of this warp function are computed.
* @param target Image of <em>n+1</em> dimensions to store the steepest descent Dimension
* <em>n</em> is used to index the parameters of the warp function. For example, the partial
* derivative of the template image intensity by parameter 2 of the warp function at pixel
* <em>(x,y)</em> is stored at position <em>(x,y,1)</em>.
*/
public static <T extends NumericType<T>> void computeSteepestDescents(
final RandomAccessibleInterval<T> gradients,
final WarpFunction warpFunction,
final RandomAccessibleInterval<T> target) {
final int n = gradients.numDimensions() - 1;
final int numParameters = warpFunction.numParameters();
final T tmp = Util.getTypeFromInterval(gradients).createVariable();
for (int p = 0; p < numParameters; ++p) {
for (int d = 0; d < n; ++d) {
final Cursor<T> gd =
Views.flatIterable(Views.hyperSlice(gradients, n, d)).localizingCursor();
for (final T t : Views.flatIterable(Views.hyperSlice(target, n, p))) {
tmp.set(gd.next());
tmp.mul(warpFunction.partial(gd, d, p));
t.add(tmp);
}
}
}
}
@Override
public RandomAccessibleInterval<T> compute(
final RandomAccessibleInterval<T> input, final RandomAccessibleInterval<T> output) {
final StructuringElementCursor<T> inStructure =
new StructuringElementCursor<T>(Views.extend(input, m_factory).randomAccess(), m_struc);
final Cursor<T> out = Views.iterable(output).localizingCursor();
double m;
while (out.hasNext()) {
out.next();
inStructure.relocate(out);
inStructure.next();
m = inStructure.get().getRealDouble();
while (inStructure.hasNext()) {
inStructure.next();
m = Math.min(m, inStructure.get().getRealDouble());
}
out.get().setReal(m);
}
return output;
}
double alignStep(final RandomAccessibleInterval<T> image) {
// compute error image = warped image - template
computeDifference(Views.extendBorder(image), currentTransform, template, error);
// compute transform parameter update
final double[] gradient = new double[numParameters];
for (int p = 0; p < numParameters; ++p) {
final Cursor<T> err = Views.flatIterable(error).cursor();
for (final T t : Views.flatIterable(Views.hyperSlice(descent, n, p)))
gradient[p] += t.getRealDouble() * err.next().getRealDouble();
}
final double[] dp = new double[numParameters];
LinAlgHelpers.mult(Hinv, gradient, dp);
// udpate transform
currentTransform.preConcatenate(warpFunction.getAffine(dp));
// return norm of parameter update vector
return LinAlgHelpers.length(dp);
}
@Test
public void defaultCollapseNumericTest() {
Img<NativeARGBDoubleType> img =
new ArrayImgFactory<NativeARGBDoubleType>()
.create(new int[] {10, 10}, new NativeARGBDoubleType());
CompositeIntervalView<NativeARGBDoubleType, NumericComposite<NativeARGBDoubleType>> il2 =
Views.collapseNumeric((RandomAccessibleInterval<NativeARGBDoubleType>) img);
CompositeIntervalView<NativeARGBDoubleType, NumericComposite<NativeARGBDoubleType>> opr =
ops.transform().collapseNumeric((RandomAccessibleInterval<NativeARGBDoubleType>) img);
assertEquals(il2.numDimensions(), opr.numDimensions());
CompositeView<NativeARGBDoubleType, NumericComposite<NativeARGBDoubleType>> il2_2 =
Views.collapseNumeric((RandomAccessible<NativeARGBDoubleType>) img, 1);
CompositeView<NativeARGBDoubleType, NumericComposite<NativeARGBDoubleType>> opr_2 =
ops.transform().collapseNumeric((RandomAccessible<NativeARGBDoubleType>) img, 1);
assertEquals(il2_2.numDimensions(), opr_2.numDimensions());
}
/**
* Extends the given {@link ImgPlus} to 5-dimensions and makes sure, that the dimension order
* suits IJ.
*
* @param img to be extended and permuted
* @return extended and permuted {@link Img}
*/
public static <T> RandomAccessibleInterval<T> extendAndPermute(final ImgPlus<T> img) {
// Create Mapping [at position one -> new index, at position 2 -> new index etc.] given: ImgPlus
// and m_mapping
// we always want to have 5 dimensions
RandomAccessibleInterval<T> extended = img;
for (int d = img.numDimensions(); d < 5; d++) {
extended = Views.addDimension(extended, 0, 0);
}
return DimSwapper.swap(extended, getInferredMapping(img));
}
@Override
public void run() {
double[] sigmas = sigmas();
@SuppressWarnings("unchecked")
Img<T> target = (Img<T>) dataset.getImgPlus();
Img<T> input = target.copy();
ExtendedRandomAccessibleInterval<T, ?> paddedInput = Views.extendMirrorSingle(input);
try {
Gauss3.gauss(sigmas, paddedInput, target);
} catch (Exception e) {
cancel(e.getMessage());
}
}
/**
* Walk through an Interval on an Img using a LocalizingCursor, localizing on every step.
*
* @param img
* @param interval
*/
protected static void localizingWalkThrough(final Img<IntType> img, final Interval interval) {
final Cursor<IntType> c = Views.interval(img, interval).localizingCursor();
final long[] pos = new long[interval.numDimensions()];
int i = 0;
while (c.hasNext()) {
c.fwd();
i += c.get().get();
c.localize(pos);
}
j = (int) pos[0] + i;
}
private Img<?> getUsableImage(final Img<ComplexType<?>> img) {
if (m_imgBuffer.size() > 0) {
return m_imgBuffer.poll();
} else {
Img tmpImg = m_imgFactory.create(img, m_resultType);
// wrap the result image with an ImgView in case the other image has an offset (i.e. minimum)
// -> if not set, an iteration order exception will occur
long[] min = new long[img.numDimensions()];
img.min(min);
return ImgView.wrap(Views.translate(tmpImg, min), tmpImg.factory());
}
}
public static final <T extends RealType<T>> void addGaussianNoiseToImage(
RandomAccessibleInterval<T> img, double sigma_noise) {
IterableInterval<T> iterImg = Views.iterable(img);
Cursor<T> lc = iterImg.localizingCursor();
double val;
T var = iterImg.firstElement().createVariable();
while (lc.hasNext()) {
lc.fwd();
val = Math.max(0, sigma_noise * ran.nextGaussian());
var.setReal(val);
lc.get().add(var);
}
}
/**
* TODO
*
* @param cumulativeMinCutoff
* @param cumulativeMaxCutoff
* @param state
* @param setupAssignments
*/
public static void initBrightness(
final double cumulativeMinCutoff,
final double cumulativeMaxCutoff,
final ViewerState state,
final SetupAssignments setupAssignments) {
final Source<?> source = state.getSources().get(state.getCurrentSource()).getSpimSource();
final int timepoint = state.getCurrentTimepoint();
if (!source.isPresent(timepoint)) return;
if (!UnsignedShortType.class.isInstance(source.getType())) return;
@SuppressWarnings("unchecked")
final RandomAccessibleInterval<UnsignedShortType> img =
(RandomAccessibleInterval<UnsignedShortType>)
source.getSource(timepoint, source.getNumMipmapLevels() - 1);
final long z = (img.min(2) + img.max(2) + 1) / 2;
final int numBins = 6535;
final Histogram1d<UnsignedShortType> histogram =
new Histogram1d<UnsignedShortType>(
Views.iterable(Views.hyperSlice(img, 2, z)),
new Real1dBinMapper<UnsignedShortType>(0, 65535, numBins, false));
final DiscreteFrequencyDistribution dfd = histogram.dfd();
final long[] bin = new long[] {0};
double cumulative = 0;
int i = 0;
for (; i < numBins && cumulative < cumulativeMinCutoff; ++i) {
bin[0] = i;
cumulative += dfd.relativeFrequency(bin);
}
final int min = i * 65535 / numBins;
for (; i < numBins && cumulative < cumulativeMaxCutoff; ++i) {
bin[0] = i;
cumulative += dfd.relativeFrequency(bin);
}
final int max = i * 65535 / numBins;
final MinMaxGroup minmax = setupAssignments.getMinMaxGroups().get(0);
minmax.getMinBoundedValue().setCurrentValue(min);
minmax.getMaxBoundedValue().setCurrentValue(max);
}
protected void generateHistogramData(DataContainer<T> container) {
double ch1BinWidth = getXBinWidth(container);
double ch2BinWidth = getYBinWidth(container);
// get the 2 images for the calculation of Pearson's
final RandomAccessibleInterval<T> img1 = getImageCh1(container);
final RandomAccessibleInterval<T> img2 = getImageCh2(container);
final RandomAccessibleInterval<BitType> mask = container.getMask();
// get the cursors for iterating through pixels in images
TwinCursor<T> cursor =
new TwinCursor<T>(
img1.randomAccess(), img2.randomAccess(), Views.iterable(mask).localizingCursor());
// create new image to put the scatter-plot in
final ImgFactory<LongType> scatterFactory = new ArrayImgFactory<LongType>();
plotImage = scatterFactory.create(new int[] {xBins, yBins}, new LongType());
// create access cursors
final RandomAccess<LongType> histogram2DCursor = plotImage.randomAccess();
// iterate over images
long[] pos = new long[plotImage.numDimensions()];
while (cursor.hasNext()) {
cursor.fwd();
double ch1 = cursor.getFirst().getRealDouble();
double ch2 = cursor.getSecond().getRealDouble();
/* Scale values for both channels to fit in the range.
* Moreover mirror the y value on the x axis.
*/
pos[0] = getXValue(ch1, ch1BinWidth, ch2, ch2BinWidth);
pos[1] = getYValue(ch1, ch1BinWidth, ch2, ch2BinWidth);
// set position of input/output cursor
histogram2DCursor.setPosition(pos);
// get current value at position and increment it
long count = histogram2DCursor.get().getIntegerLong();
count++;
histogram2DCursor.get().set(count);
}
xBinWidth = ch1BinWidth;
yBinWidth = ch2BinWidth;
xLabel = getLabelCh1();
yLabel = getLabelCh2();
xMin = getXMin(container);
xMax = getXMax(container);
yMin = getYMin(container);
yMax = getYMax(container);
}
public static <T extends Type<T> & Comparable<T>> int findLocalMaximaNeighborhood2(
final RandomAccessibleInterval<T> img) {
// final ArrayList< Point > maxima = new ArrayList< Point >();
int nMaxima = 0;
final Cursor<T> center =
Views.iterable(Views.interval(img, Intervals.expand(img, -1))).localizingCursor();
final LocalNeighborhood2<T> neighborhood = new LocalNeighborhood2<T>(img, center);
final LocalNeighborhoodCursor2<T> nc = neighborhood.cursor();
A:
while (center.hasNext()) {
final T t = center.next();
neighborhood.updateCenter(center);
nc.reset();
while (nc.hasNext()) {
final T n = nc.next();
if (n.compareTo(t) > 0) continue A;
}
// maxima.add( new Point( center ) );
++nMaxima;
}
return nMaxima;
}
/**
* Computes a Gaussian convolution in-place (temporary imgs are necessary) with float precision on
* an entire {@link Img}
*
* @param sigma - the sigma for the convolution
* @param img - the img {@link Img} that will be convolved in place
*/
public static <T extends RealType<T>> void inFloatInPlace(
final double[] sigma,
final Img<T> img,
final OutOfBoundsFactory<FloatType, RandomAccessibleInterval<FloatType>> outofbounds) {
GaussFloat gauss = null;
try {
if (FloatType.class.isInstance(img.firstElement())) {
@SuppressWarnings({"rawtypes", "unchecked"})
final Img<FloatType> img2 = (Img) img;
gauss =
new GaussFloat(
sigma,
Views.extend(img2, outofbounds),
img2,
img2,
new Point(sigma.length),
img2.factory().imgFactory(new FloatType()));
} else {
final RandomAccessibleInterval<FloatType> rIn =
new WriteConvertedIterableRandomAccessibleInterval<T, FloatType, Img<T>>(
img, new RealFloatSamplerConverter<T>());
gauss =
new GaussFloat(
sigma,
Views.extend(rIn, outofbounds),
img,
rIn,
new Point(sigma.length),
img.factory().imgFactory(new FloatType()));
}
} catch (final IncompatibleTypeException e) {
System.out.println(e);
return;
}
gauss.call();
}
private static final <R extends RealType<R>> Img<R> processReal(
final Img<R> img, final float[] m, final Mode mode, final OutOfBoundsFactory<R, Img<R>> oobf)
throws Exception {
final InterpolatorFactory<R, RandomAccessible<R>> inter;
switch (mode) {
case LINEAR:
inter = new NLinearInterpolatorFactory<R>();
break;
case NEAREST_NEIGHBOR:
inter = new NearestNeighborInterpolatorFactory<R>();
break;
default:
throw new IllegalArgumentException("Scale: don't know how to scale with mode " + mode);
}
final ImageTransform<R> transform;
final ExtendedRandomAccessibleInterval<R, Img<R>> imgExt = Views.extend(img, oobf);
if (2 == img.numDimensions()) {
// Transform the single-plane image in 2D
AffineModel2D aff = new AffineModel2D();
aff.set(m[0], m[4], m[1], m[5], m[3], m[7]);
transform = new ImageTransform<R>(imgExt, aff, (InterpolatorFactory) inter);
} else if (3 == img.numDimensions()) {
// Transform the image in 3D, or each plane in 2D
if (m.length < 12) {
throw new IllegalArgumentException(
"Affine transform in 3D requires a matrix array of 12 elements.");
}
AffineModel3D aff = new AffineModel3D();
aff.set(m[0], m[1], m[2], m[3], m[4], m[5], m[6], m[7], m[8], m[9], m[10], m[11]);
transform = new ImageTransform<R>(imgExt, aff, (InterpolatorFactory) inter);
// Ensure Z dimension is not altered if scaleZ is 1:
if (Math.abs(m[10] - 1.0f) < 0.000001 && 0 == m[8] && 0 == m[9]) {
long[] d = transform.getNewImageSize();
d[2] = img.dimension(2); // 0-based: '2' is the third dimension
transform.setNewImageSize(d);
}
} else {
throw new Exception("Affine transform: only 2D and 3D images are supported.");
}
if (!transform.checkInput() || !transform.process()) {
throw new Exception("Could not affine transform the image: " + transform.getErrorMessage());
}
return transform.getResult();
}
public static final <T extends RealType<T>> void addGaussianSpotToImage(
RandomAccessibleInterval<T> img, double[] params) {
IterableInterval<T> iterImg = Views.iterable(img);
Cursor<T> lc = iterImg.localizingCursor();
int nDims = img.numDimensions();
double[] position = new double[nDims];
double val;
T var = iterImg.firstElement().createVariable();
while (lc.hasNext()) {
lc.fwd();
lc.localize(position);
val = gaussian.val(position, params);
var.setReal(val);
lc.get().add(var);
}
}
private void computeDataMinMax(final RandomAccessibleInterval<? extends RealType<?>> img) {
// FIXME: Reconcile this with DefaultDatasetView.autoscale(int). There is
// no reason to hardcode the usage of ComputeMinMax twice. Rather, there
// should be a single entry point for obtain the channel min/maxes from
// the metadata, and if they aren't there, then compute them. Probably
// Dataset (not DatasetView) is a good place for it, because it is metadata
// independent of the visualization settings.
DataRange range = autoscaleService.getDefaultRandomAccessRange(img);
dataMin = range.getMin();
dataMax = range.getMax();
// System.out.println("IN HERE!!!!!!");
// System.out.println(" dataMin = " + dataMin);
// System.out.println(" dataMax = " + dataMax);
@SuppressWarnings({"unchecked", "rawtypes"})
Iterable<T> iterable =
Views.iterable((RandomAccessibleInterval<T>) (RandomAccessibleInterval) img);
BinMapper1d<T> mapper = new Real1dBinMapper<T>(dataMin, dataMax, 256, false);
Histogram1d<T> histogram = new Histogram1d<T>(iterable, mapper);
if (bundle == null) {
bundle = new HistogramBundle(histogram);
} else {
bundle.setHistogram(histogram);
}
bundle.setDataMinMax(dataMin, dataMax);
// bundle.setLineSlopeIntercept(1, 0);
log.debug("computeDataMinMax: dataMin=" + dataMin + ", dataMax=" + dataMax);
// force a widget refresh to see new Hist (and also fill min and max fields)
// NOPE. HistBundle is unchanged. Only internals are. So no
// refresh called. Note that I changed InteractiveCommand::update() to
// always setValue() and still this did not work. !!!! Huh?
// update(getInfo().getMutableInput("bundle", HistogramBundle.class),
// bundle);
// NOPE
// getInfo().getInput("bundle", HistogramBundle.class).setValue(this,
// bundle);
// NOPE
// getInfo().setVisible(false);
// getInfo().setVisible(true);
// NOPE
// getInfo().getMutableInput("bundle",HistogramBundle.class).initialize(this);
// NOPE
// getInfo().getMutableInput("bundle",HistogramBundle.class).callback(this);
}
private boolean processSlice(
final RandomAccessibleInterval<FloatType> src,
final RandomAccessibleInterval<FloatType> dx,
final RandomAccessibleInterval<FloatType> dy) {
// Gaussian filter.
final ExtendedRandomAccessibleInterval<FloatType, RandomAccessibleInterval<FloatType>>
extended = Views.extendMirrorSingle(src);
try {
Gauss3.gauss(new double[] {sigma, sigma}, extended, src, numThreads);
} catch (final IncompatibleTypeException e) {
errorMessage = BASE_ERROR_MSG + "Incompatible types: " + e.getMessage();
e.printStackTrace();
return false;
}
// Derivatives
PartialDerivative.gradientCentralDifference(extended, dx, 0);
PartialDerivative.gradientCentralDifference(extended, dy, 1);
return true;
}
private static void adjustForOSEM(
final HashMap<ViewId, RandomAccessibleInterval<FloatType>> weights,
final WeightType weightType,
final double osemspeedup) {
if (osemspeedup == 1.0) return;
if (weightType == WeightType.PRECOMPUTED_WEIGHTS
|| weightType == WeightType.WEIGHTS_ONLY
|| weightType == WeightType.LOAD_WEIGHTS) {
for (final RandomAccessibleInterval<FloatType> w : weights.values()) {
for (final FloatType f : Views.iterable(w))
f.set(
Math.min(
1, f.get() * (float) osemspeedup)); // individual contribution never higher than 1
}
} else if (weightType == WeightType.NO_WEIGHTS) {
for (final RandomAccessibleInterval<FloatType> w : weights.values()) {
final RandomAccess<FloatType> r = w.randomAccess();
final long[] min = new long[w.numDimensions()];
w.min(min);
r.setPosition(min);
r.get()
.set(
Math.min(
1,
r.get().get()
* (float) osemspeedup)); // individual contribution never higher than 1
}
} else if (weightType == WeightType.VIRTUAL_WEIGHTS) {
for (final RandomAccessibleInterval<FloatType> w : weights.values())
((NormalizingRandomAccessibleInterval<FloatType>) w).setOSEMspeedup(osemspeedup);
} else {
throw new RuntimeException(
"Weight Type: "
+ weightType.name()
+ " not supported in ProcessForDeconvolution.adjustForOSEM()");
}
}
public Align(final RandomAccessibleInterval<T> template, final ImgFactory<T> factory) {
this.template = template;
final T type = Util.getTypeFromInterval(template);
n = template.numDimensions();
warpFunction = new AffineWarp(n);
numParameters = warpFunction.numParameters();
currentTransform = new AffineTransform(n);
final long[] dim = new long[n + 1];
for (int d = 0; d < n; ++d) dim[d] = template.dimension(d);
dim[n] = n;
final Img<T> gradients = factory.create(dim, type);
gradients(Views.extendBorder(template), gradients);
dim[n] = numParameters;
descent = factory.create(dim, type);
computeSteepestDescents(gradients, warpFunction, descent);
Hinv = computeInverseHessian(descent);
error = factory.create(template, type);
}
