Anti-Aliasing / How to reconstruct a zoomed image - DICOM

This is a discussion on Anti-Aliasing / How to reconstruct a zoomed image - DICOM ; Hello all, I am developing a DICOM Viewer application and I have the following question on regards to the reconstruction of a zoomed image: My viewer application opens a native DICOM image -as is-. This means that our 100% represents ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: Anti-Aliasing / How to reconstruct a zoomed image

  1. Anti-Aliasing / How to reconstruct a zoomed image

    Hello all,

    I am developing a DICOM Viewer application and I have the following
    question on regards to the reconstruction of a zoomed image:

    My viewer application opens a native DICOM image -as is-. This means
    that our 100% represents the image in the same size as it was taken
    from the CT or CR. Then, when we apply a zoom tool to enlarge up to
    200% the DICOM image automatically show a "pixeled" image, and the loss
    in sharpness is evident. Moreover, when re-zooming up to 300% the image
    gets completely pixeled.

    Is there a way that we could rebuild the sharpness of the zoomed image
    in every zooming step we apply to the image? Is there an API or any
    other algorithm to reconstruct a pixeled image?

    I will really appreciate your comments. We could take any advice as
    good, even if it is a developed API application ($) or any line of code
    for shareware or freeware.

    Thanks in advance for your help.

    Regards,

    Lic. Erasmo A. González Pérez
    IMSS-VistA / Imaginología
    IMSS-VistA / Nota de Enfermería
    ______________________________________
    Tokio 80, 4o. Piso
    erasmo.gonzalez@softtek.com
    Tel. 5238-2700 ext 12668 / 12642
    Cel. Softtek 044-55-1847-2161


  2. Re: Anti-Aliasing / How to reconstruct a zoomed image

    Hola Erasmo!

    What you need to do is use interpolation to zoom the image, there are
    several methods, linear interpolation is easy to implement and fast.
    Bi-cubic interpolation would give you better quality but takes more
    processing power.

    If you need any help don't doubt to contact me,

    Regards,
    Rafael Sanguinetti
    CharruaSoft.com

    erasmo.gonzalez@softtek.com ha escrito:

    > Hello all,
    >
    > I am developing a DICOM Viewer application and I have the following
    > question on regards to the reconstruction of a zoomed image:
    >
    > My viewer application opens a native DICOM image -as is-. This means
    > that our 100% represents the image in the same size as it was taken
    > from the CT or CR. Then, when we apply a zoom tool to enlarge up to
    > 200% the DICOM image automatically show a "pixeled" image, and the loss
    > in sharpness is evident. Moreover, when re-zooming up to 300% the image
    > gets completely pixeled.
    >
    > Is there a way that we could rebuild the sharpness of the zoomed image
    > in every zooming step we apply to the image? Is there an API or any
    > other algorithm to reconstruct a pixeled image?
    >
    > I will really appreciate your comments. We could take any advice as
    > good, even if it is a developed API application ($) or any line of code
    > for shareware or freeware.
    >
    > Thanks in advance for your help.
    >
    > Regards,
    >
    > Lic. Erasmo A. González Pérez
    > IMSS-VistA / Imaginología
    > IMSS-VistA / Nota de Enfermería
    > ______________________________________
    > Tokio 80, 4o. Piso
    > erasmo.gonzalez@softtek.com
    > Tel. 5238-2700 ext 12668 / 12642
    > Cel. Softtek 044-55-1847-2161



  3. Re: Anti-Aliasing / How to reconstruct a zoomed image

    Hi Erasmo,

    There are several open source image manipulation libraries. ImageMagick
    (www.imagemagick.org) and its clone GraphicsMagick
    (www.graphicsmagick.org) are widely known. Both are highly portable and
    supported in several programing languages.

    Check the available resizing algorithms here:
    http://www.imagemagick.org/api/magic...ckCommentImage

    Note that most graphics libraries (including the above mentioned) are
    designed for computer graphics and publishing applications, not for
    medical imaging. So, they are aiming at 'visual enhancment' of an image
    rather than 'correctness'. Some algorithms may add incorrect details,
    like halos, which may lead to incorrect diagnostic. You should have
    experienced radiologist extensively test the chosen algorithm with
    different studies before it is used in actual diagnostic.

    Another note is that you also should use good algorithm for
    downsampling too (zoom out) as it greatly enhances the small image (in
    contrast to removing rows and columns).

    These libraries may also have several important functions for your
    application, like importing and exporting other image formats, changing
    'Levels' ('Window' in medical terminology).

    Hope this helps,

    Mohammad Alhashash


+ Reply to Thread