[Insight-users] (no subject): Image Quality : Deblurring ?
Luis Ibanez
luis . ibanez at kitware . com
Thu, 25 Sep 2003 13:53:06 -0400
Hi Rashmi,
Thanks for your additional details.
I would suggest you to try any/all of the
anisotropic diffusion filters:
- CurvatureFlowImageFilter
- GradientAnisotropicDiffusionImageFilter
- CurvatureAnistropicDiffusion
- BilateralImageFilter
They may help redude the fuzziness of your image.
You will find them described in the SoftwareGuide
http://www . itk . org/ItkSoftwareGuide . pdf
Section: 6.5.2, pdf-page 173.
Note that if the actual width of the bars if of
significance for your processing you will have
to do some validation in order to figure out if
the processing will report bars that are wider
or narrower than the original ones.
---
Now that, if the blurring is due to fast linear
motion during the acquisition of the image, you may
want to consider using deconvolution methods, a search
on IEEE-PAMI will give you some useful references.
There is also a signficant amount of work on deblurring
images for microscopy using deconvolution methods. They
may be applicable to your problem too.
Regards,
Luis
------------------
Rashmi Ghai wrote:
> Hi Luis ,
> I'm dealing with 2D images , in grey scale, for eg images of bar codes.
> My input is blurred and unfocussed images which have to be processed so
> as to produce clear images which can be decoded.
> Guassian and laplacian filters are not producing much clarity.What other
> sequence of filters would solve this problem?
> -Rashmi
>
> -----Original Message-----
> From: Luis Ibanez [mailto:luis . ibanez at kitware . com]
> Sent: Thursday, September 25, 2003 12:00 AM
> To: Rashmi Ghai
> Cc: insight-users at itk . org
> Subject: Re: [Insight-users] (no subject): Image Quality : Deblurring ?
>
>
> Hi Rashmi,
>
> "Image quality" is a very ambiguous term.
>
> Could you please provide more details
> about your problem ?
>
> What is the modality of your image ?
>
> You mention focus,... are you dealing with
> microscopy images ? have you considered
> deconvolution ?
>
> ----
>
> With the very limited information that you
> provided, the only thing that comes to mind
> is to suggest you to use the adaptive constrast
> enhancement filter.
> http://www . itk . org/Insight/Doxygen/html/classitk_1_1AdaptiveHistogramEqu
> alizationImageFilter.html
>
>
> This has mostly been used in mamography for
> improving the visibility of anatomical structures
> as perceived by human observers.
>
>
>
> Luis
>
>
>
> ----------------------
>
> Rashmi Ghai wrote:
>
>>hi everyone
>>
>>I'm new to using ITK and my task is as follows
>>
>>Given an image , i have to improve its quality and
>>
>>majorly improve its focus.
>>
>>Please suggest what filters and in what order can/should be applied?
>>
>>Thanks
>>
>>-rashmi
>>
>
>
>
>
>
>
>
>