Posted by
Herbie on
Oct 17, 2019; 9:44am
URL: http://imagej.273.s1.nabble.com/how-does-ImageJ-measure-circular-ROI-tp5022481p5022549.html
Good day Charitra,
here are (again) some ideas:
1. It is always a good idea to mention, if a topic is cross-posted, so
that those interested in it are able to follow comments and solutions.
(I've mentioned the cross-post in my first reply to you, here on the
list. Marcel aka Bio7 did the same today on the Forum.)
2. Meanwhile and as Michael has mentioned, you've realized that ImageJ
uses the "floor"-rule when dealing with circular ROIs, i.e. the next
lowest integer.
3. As mentioned in my first reply, I recommend to scale the images using
an adequate interpolation scheme. When doing so, the problem of
"percentage pixels" can be reduced to integer pixels. Please note that
interpolated scaling doesn't create information but can circumvent the
problem of "percentage pixels".
Regards
Herbie
::::::::::::::::::::::::::::::::::::::
Am 17.10.19 um 11:03 schrieb Charitra:
> Hello Herbie and Michael
>
> Thank you for your replies.
>
> Herbie, I had posted my question on both the mailing list and the forum on
> the same day. Not sure what you mean by reposting. I'm not very familiar
> with how the forum and mailing list work, sorry.
>
> But coming back, this was what I suspected - that ImageJ does not care about
> sub-pixels. I am not interested in filling sub-pixel oval ROIs; I want to
> measure them. So I just used 'fill' in a way to visualize which pixels were
> going into the measurement. But coming back to Michael's answer, yes, the
> values in the ROI managers are always integers, but if I 'specify' a ROI and
> add it to the ROI manager, although the values are an integer, the sub-pixel
> nature of the ROI is maintained. Or was this a wrong observation of mine?
>
> Also, if there is no way to measure just 40% or 50% of the pixel, then I am
> afraid I cannot use the oval ROIs. I have switched to using square ROIs
> instead, as I cannot afford to do guesswork on these intensity measurements.
> Moreover, in oval ROIs, the number of pixels taken into consideration is not
> clear and I need to know exactly which and how many pixels are taken into
> account. So I think I am going to continue with rectangular ROIs. Sorry for
> the delay in replying, and have a nice day!
>
> Charitra.
>
>
>
> --
> Sent from:
http://imagej.1557.x6.nabble.com/>
> --
> ImageJ mailing list:
http://imagej.nih.gov/ij/list.html>
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html