Field Detail

• patches

  public final List<Patch> patches

• finalBox

  public final Rectangle finalBox

• backgroundValue

  public final int backgroundValue

• scale

  public final double scale

• largest_possibly_needed_area

  public final double largest_possibly_needed_area

• max_possible_area

  public final double max_possible_area

• scaleUP

  public final double scaleUP

• loader

  protected final Loader loader

Constructor Detail

• ExportBestFlatImage

  public ExportBestFlatImage(List<Patch> patches,
                             Rectangle finalBox,
                             int backgroundValue,
                             double scale)

  Class to manage the creation of an ImageProcessor containing a flat 8-bit or RGB image for use in e.g. extracting features, and optionally do so along with its alpha mask as another ImageProcessor. If the image is smaller than 0.5 GB, then the AWT system in Loader.getFlatAWTImage will be used, with the quality flag as true. Otherwise, mipmaps will be used to generate an image with ExportUnsignedByte.makeFlatImage. If mipmaps are not available, then an up to 2GB image will be generated with ExportUnsignedShort (which respects pixel intensity mappings) and then Gaussian-downsampled to the requested dimensions. This method is as safe as it gets, regarding the many caveats of alpha masks, min-max, and the 2GB array indexing limit of java-8.

  Parameters:

  patches -

  finalBox -

  backgroundValue -

  scale -

Method Detail

• canUseAWTImage

  public boolean canUseAWTImage()

  Returns:

  Whether an AWT image can be used: the requested area must be smaller than 0.5 GB, so that with the quality flag, the interim image is smaller than 2 GB (2x larger on the side).

• isSmallerThan2GB

  public boolean isSmallerThan2GB()

  Returns:

  Whether the requested image fits in an array up to 2 GB in size.

• printInfo

  public void printInfo()

• convertToFloat

  protected ij.process.FloatProcessor convertToFloat(ij.process.ShortProcessor sp)

• computeSigma

  public double computeSigma(int width,
                             int height)

• gaussianDownsampled

  protected ij.process.FloatProcessor gaussianDownsampled(ij.process.FloatProcessor ip)

  Gaussian-downsample to the target dimensions

  Parameters:

  ip -

  Returns:

• createAWTImage

  protected Image createAWTImage(int type,
                                 Color background)

  Create a java.awt.Image using the Loader.getFlatAWTImage method, using 'true' for the quality flag, which means than an image 2x larger on each side will be generated and then scaled down by area averaging. Uses mipmaps, which are Gaussian-downsampled already.

  Parameters:

  type - Either ImagePlus.GRAY8 (TYPE_BYTE_INDEXED) or ImagePlus.COLOR_RGB (TYPE_INT_ARGB)

  background -

• createAWTImage

  protected Image createAWTImage(int type)

  Create a java.awt.Image using the Loader.getFlatAWTImage method, using 'true' for the quality flag, which means than an image 2x larger on each side will be generated and then scaled down by area averaging. Uses mipmaps.

  Parameters:

  type - Either ImagePlus.GRAY8 (TYPE_BYTE_INDEXED) or ImagePlus.COLOR_RGB (TYPE_INT_ARGB)

• makeFlatColorImage

  public Pair<ij.process.ColorProcessor,ij.process.ByteProcessor> makeFlatColorImage()

• makeFlatGrayImage

  public ij.process.ByteProcessor makeFlatGrayImage()

  Returns:

  null when the dimensions make the array larger than 2GB, or the image otherwise.

• makeFlatGrayImageAndAlpha

  public Pair<ij.process.ByteProcessor,ij.process.ByteProcessor> makeFlatGrayImageAndAlpha()

  Returns:

  Return null when dimensions make the array larger than 2GB.

• makeFlatFloatGrayImageAndAlpha

  public Pair<ij.process.FloatProcessor,ij.process.FloatProcessor> makeFlatFloatGrayImageAndAlpha()

  While the data is in the 8-bit range, the format is as a FloatProcessor.