Image J 16 bit Signed Buffered Image

Question

So my question is how to get 16 bit bufferedImage from ij.ImagePlus...? If am trying to get using ShortProcessor it change my signed image to unsigned so i am not getting original image...Thanks in advance can any one provide solution.

enter image description here

How ImageJ display 16 bit signed image in their viewer..and we only get 8 bit bufferedImage or 16 bit unsigned bufferedImage So how can i get 16 bit signed BufferedImage..?

How do you get a signed 16-bit image in an `ImagePlus`? To my knowledge, ImageJ only supports _unsigned_ 16-bit images. What does [`ImagePlus#getType()`](http://jenkins.imagej.net/job/ImageJ1-javadoc/javadoc/ij/ImagePlus.html#getType%28%29) return? — Jan Eglinger, Apr 24 '14 at 13:53
@JanEglinger ImageJ supports calibration functions though, which can be used to "simulate" signed 16-bit data. What kamlesh0606 could do is get the unsigned BufferedImage back, then convert it to a signed BufferedImage by looping over the pixels. I'm too busy at the moment to write this up as a proper answer, though... — ctrueden, Apr 25 '14 at 03:07
Hey I also facing the same problem...can any one give answer..? — Jay Thakkar, Apr 25 '14 at 10:06

score 1 · Answer 1 · answered Apr 27 '14 at 03:49

ImageJ can represent a signed 16-bit type using a special Calibration function. The isSigned16Bit() method indicates when that specific calibration function is in use—it is a linear m*x+b calibration where m=1 and b=-32768; this can be seen in the ImageJ source code.

ImageJ provides a way to obtain a BufferedImage from an ImagePlus via the getImage() method. However, this always returns an 8-bit BufferedImage.

So the next approach is to create your own BufferedImage with type DataBuffer.TYPE_SHORT, which wraps the same short[] array that backs the original ImagePlus object. Unfortunately, due to ImageJ's internal representation of signed 16-bit data, the values will be off by a constant offset of 32768—e.g., a raw value of -444 will be stored in ImageJ's short[] array as 32324. Due to this fact, you must manually adjust all your values before wrapping as a BufferedImage.

Here is some example code:

import io.scif.gui.AWTImageTools;
...

final ImagePlus imp =
    IJ.openImage("http://imagej.net/images/ct.dcm.zip");

// get pixels array reference
final short[] pix = (short[]) imp.getProcessor().getPixels();
final int w = imp.getWidth();
final int h = imp.getHeight();
final boolean signed = imp.getCalibration().isSigned16Bit();

if (signed) {
    // adjust raw pixel values
    for (int i=0; i<pix.length; i++) {
        pix[i] -= 32768;
    }
}

// convert to BufferedImage
final BufferedImage image = AWTImageTools.makeImage(pix, w, h, signed);

For the actual conversion of short[] to BufferedImage, this code makes use of the SCIFIO library's AWTImageTools.makeImage utility methods. SCIFIO is included with the Fiji distribution of ImageJ. Alternately, it is only a few lines of code which would be easy to copy and paste from the routine in question.

i am doing same procedure but not getting original image.still not get clear image(in contrast with gray but in real require black) — Kamlesh, Apr 28 '14 at 04:54
Inspect the resultant `BufferedImage`'s pixel values to be certain, but my guess is that the image colors are simply not scaled the way you expect, because it is a 16-bit image. Java does not autoscale `BufferedImage`s when painting them. — ctrueden, Apr 28 '14 at 20:55

Image J 16 bit Signed Buffered Image

1 Answers1