[vtkusers] bit wise operations for vtkImageData
Jonathan Morra
jonmorra at gmail.com
Mon Mar 28 17:23:09 EDT 2011
So after some more work, I figured out that's it definitely the setCharArray
method that's not working.
After testing I have realized that this works but is very slow (too slow to
be usable)
private void setCharArray(vtkImageData image, char[] array) {
vtkCharArray ca = new vtkCharArray();
ca.Initialize();
ca.SetNumberOfValues(array.length);
//ca.SetJavaArray(array);
int i=0;
for (char data : array) {
ca.SetValue(i, data);
++i;
}
ca.Modified();
ca.SetName("ImageScalars");
image.GetPointData().SetScalars(ca);
image.Update();
}
and this does not work (but it is fast about it's not working)
private void setCharArray(vtkImageData image, char[] array) {
vtkCharArray ca = new vtkCharArray();
ca.Initialize();
ca.SetNumberOfValues(array.length);
ca.SetJavaArray(array);
/*int i=0;
for (char data : array) {
ca.SetValue(i, data);
++i;
}*/
ca.Modified();
ca.SetName("ImageScalars");
image.GetPointData().SetScalars(ca);
image.Update();
}
Why isn't SetJavaArray() working?
On Mon, Mar 28, 2011 at 2:17 PM, David Gobbi <david.gobbi at gmail.com> wrote:
> I can't see anything that is obviously wrong, but my experience
> with Java is limited.
>
> - David
>
>
> On Mon, Mar 28, 2011 at 2:42 PM, Jonathan Morra <jonmorra at gmail.com>
> wrote:
> > After playing around a little bit, I have come up with the following
> that's
> > not currently working. Could you please advise me why this isn't
> working.
> > Thanks
> > put() {
> > // Organ data contains a list of char[]
> > final char[] storedCharArray = organData.get(imageNumber);
> > final char[] binaryCharArray = getCharArray(binaryOrgan);
> > int i=0;
> > for (char binaryData : binaryCharArray) {
> > char storedData = storedCharArray[i];
> > if (binaryData == 0)
> > storedData &= ~(1 << bitNumber);
> > else
> > storedData |= (1 << bitNumber);
> > storedCharArray[i] = storedData;
> > ++i;
> > }
> > get() {
> > vtkImageData binaryImage = // Get blank image of correct
> dimensions,
> > orientation, spacing, etc.
> > final char[] storedCharArray = organData.get(imageNumber);
> > final char[] binaryCharArray = getCharArray(binaryImage);
> > int i=0;
> > for (char storedData : storedCharArray) {
> > binaryCharArray[i] = (char)((storedData >> bitNumber) & 1);
> > ++i;
> > }
> > setCharArray(binaryImage, binaryCharArray);
> > private void setCharArray(vtkImageData image, char[] array) {
> >
> > ((vtkCharArray)image.GetPointData().GetScalars()).SetJavaArray(array);
> > }
> > private char[] getCharArray(vtkImageData image) {
> > vtkPointData points = image.GetPointData();
> > vtkCharArray ca = (vtkCharArray)points.GetScalars();
> > char[] chars = ca.GetJavaArray();
> > return chars;
> > }
> >
> > On Fri, Mar 25, 2011 at 12:48 PM, David Gobbi <david.gobbi at gmail.com>
> wrote:
> >>
> >> In python it is possible to access the VTK data arrays directly via
> >> python's buffer interface, so that's what I use when I want to quickly
> >> move pixel data back and forth from python to VTK. It allows me to
> >> use a VTK array (including the one that the image uses to store its
> >> pixels) as if it was a native python array.
> >>
> >> I have a feeling that the only way you will be able to rapidly do VTK
> >> pixel operations of any kind from Java, is if someone adds a similar
> >> feature to the Java wrappers.
> >>
> >> - David
> >>
> >>
> >> On Fri, Mar 25, 2011 at 1:31 PM, Jonathan Morra <jonmorra at gmail.com>
> >> wrote:
> >> > Your understanding is correct, however I cannot
> >> > use GetScalarComponentAsDouble and SetScalarComponentAsDouble because
> >> > they
> >> > are way too slow in Java. Do you have any other suggestions for how
> to
> >> > do
> >> > this or will I have to write it myself in C++?
> >> >
> >> > On Fri, Mar 25, 2011 at 12:29 PM, David Gobbi <david.gobbi at gmail.com>
> >> > wrote:
> >> >>
> >> >> Hi Jonathan,
> >> >>
> >> >> If I understand correctly, you have several binary images with
> exactly
> >> >> the same dimensions. And you want to collapse them into a single
> >> >> image, by storing a bitfield in the pixels of that image. So:
> >> >>
> >> >> To set bit "i" in pixel "xyz" (C++ code):
> >> >> int a = int(image.GetScalarComponentAsDouble(x, y, z, 0));
> >> >> a |= (1 << i);
> >> >> image.SetScalarComponentAsDouble(x, y, z, a);
> >> >>
> >> >> To clear bit "i" in pixel "xyz":
> >> >> int a = int(image.GetScalarComponentAsDouble(x, y, z, 0));
> >> >> a &= ~(1 << i);
> >> >> image.SetScalarComponentAsDouble(x, y, z, a);
> >> >>
> >> >> To test bit "i" in pixel "xyz":
> >> >> int a = int(image.GetScalarComponentAsDouble(x, y, z, 0));
> >> >> return ((a >> i) & 1);
> >> >>
> >> >> But I'm not sure if this is what you are trying to do. I didn't
> fully
> >> >> understand your pseudocode.
> >> >>
> >> >> - David
> >> >>
> >> >>
> >> >>
> >> >> On Fri, Mar 25, 2011 at 1:08 PM, Jonathan Morra <jonmorra at gmail.com>
> >> >> wrote:
> >> >> > I'm in Java and storing a large number of binary vtkImageData's. I
> >> >> > know
> >> >> > this is inefficient, and am searching for a better way to store
> them.
> >> >> > Right
> >> >> > now I have a hash of the vtkImageData's and a get/put function.
> >> >> > Basically I
> >> >> > want to mimic this get/put but with better storage. I was thinking
> >> >> > one
> >> >> > way
> >> >> > to do this is to use bitwise logical operations to store the
> >> >> > information
> >> >> > in
> >> >> > the binary masks. For instance, if we have 2 binary images, then
> we
> >> >> > could
> >> >> > store that information in 1 vtkImageData using the following pseudo
> >> >> > code.
> >> >> > private static final int IMAGE1_BIT_CHANNEL = 0;
> >> >> > private static final int IMAGE2_BIT_CHANNEL = 1;
> >> >> > private vtkImageData storedImage;
> >> >> > vtkImageData get(String image) {
> >> >> > int channel = image eq "image1" ? IMAGE1_BIT_CHANNEL :
> >> >> > IMAGE2_BIT_CHANNEL;
> >> >> > vtkImageData return = new vtkImageData();
> >> >> > foreach (pixel in storedImage)
> >> >> > if (pixel at bit channel)
> >> >> > return[pixel] = 1;
> >> >> > else
> >> >> > return[pixel] = 0;
> >> >> > return return;
> >> >> > }
> >> >> > void put(String image, vtkImageData binaryImage) {
> >> >> > int channel = image eq "image1" ? IMAGE1_BIT_CHANNEL :
> >> >> > IMAGE2_BIT_CHANNEL;
> >> >> > foreach (pixel in binaryImage)
> >> >> > if (pixel)
> >> >> > storedImage at bit in channel = 1;
> >> >> > else
> >> >> > storedImage at bit in channel = 0;
> >> >> > }
> >> >> > This could be extended easily for 8 channels for a char image for
> >> >> > instance.
> >> >> > This operation would have to be very fast though cause it is done
> >> >> > often
> >> >> > on
> >> >> > the UI thread.
> >> >> > 1. Is there a way to do this in VTK with Java?
> >> >> > 2. Is this the best scheme for accomplishing my goal?
> >> >> > 3. Is there a better scheme for doing this?
> >> >> > Thanks.
> >> >> > _______________________________________________
> >> >> > Powered by www.kitware.com
> >> >> >
> >> >> > Visit other Kitware open-source projects at
> >> >> > http://www.kitware.com/opensource/opensource.html
> >> >> >
> >> >> > Please keep messages on-topic and check the VTK FAQ at:
> >> >> > http://www.vtk.org/Wiki/VTK_FAQ
> >> >> >
> >> >> > Follow this link to subscribe/unsubscribe:
> >> >> > http://www.vtk.org/mailman/listinfo/vtkusers
> >> >> >
> >> >> >
> >> >
> >> >
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20110328/f41994c0/attachment.htm>
More information about the vtkusers
mailing list