Re: histogram analysis

• Previous message: Dave Martindale: "Re: Q: Which 2 wavelengths give unique colour in CIE-Yxy space?"
• In reply to: Dr. Peter Boettcher: "histogram analysis"
• Next in thread: Norberto Malpica: "Re: histogram analysis"
• Messages sorted by: [ date ] [ thread ]

Date: Wed, 16 Feb 2005 00:42:51 +0100

you can always normalize the pixel values of the images that you get within
a certain range, preferably between 0 and 255 and then make a histogram
comparison. This is a first simple step, that you can test and see if it
gives satisfactory results.

if you could provide the group with links to a couple of sample photos, i
believe more suggestions would be provided by the readers...

regards

Dimitrios

"Dr. Peter Boettcher" <boettcher@kleintierklinik.uni-leipzig.de> wrote in
message news:cut7ep$gf1$3@news1.uni-leipzig.de...
> We would like to validate a method which provides a quantitative
> measurement reflecting the individual brain volume (cerebrum) in
> relation to a control group of sound dogs. Looking at the images,
> subjectively we make the diagnosis in evaluating the amount of cerebro
> spinal fluid (CSF)in contrast to the mass of the cerebrum. In diseased
> dogs one find a lot of more CSF around and "within" the cerebrum than in
> normal dogs. CSF is more on the white side of the histogram.
>
> We would like to investigate this impression in one mid-sagittal
> T2-slice of each dog. In the next step we outline the cerebrum by hand
> (one slice only) and here starts our problem. Is there any one who knows
> of any validated method to compare those ROIs in term of CSF content
> (without segmentation more in term of frequency analysis or histogram
> comparison!). We thought measuring the histogram of each cerebrum might
> reflect the amount of CSF compared to brain volume (histogram
> comparison). But up to now we do not know how to do such a comparison
> the right way and I am not sure whether it would be the right way,
> because the images from one to the next patient can differ in term of
> contrast and brightness (which would be reflected in a different
> histogram) but anatomically they are both of the same shape (no
shrinkage).
>
> Regards, Peter.

• Previous message: Dave Martindale: "Re: Q: Which 2 wavelengths give unique colour in CIE-Yxy space?"
• In reply to: Dr. Peter Boettcher: "histogram analysis"
• Next in thread: Norberto Malpica: "Re: histogram analysis"
• Messages sorted by: [ date ] [ thread ]