Re: I FOUND the little SQUARES (fwd)

[Johannes Schindelin]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi list,

Sorry, was sending from the wrong address, so the list rejected it.
Therefore I am resending.

But this gives me the opportunity to state my point more clearly:

        I do not take exception with displaying pixels as squares if that
        helps analyze the data.

        I do take exception with people pretending that pixels _are_
        square. They are not.

I saw my share of bad analyses that were based on that false notion that
pixels have an exact square-shaped extent (or for that matter, voxels that
have an exact cuboid-shaped extent). For example, volume estimates of
reconstructed objects of interest were wildly off because of that (and
were reported without any indication of the expected precision), it was
not even funny.

Ciao,
Johannes

---------- Forwarded message ----------
Date: Tue, 17 Apr 2012 10:21:04 -0500 (CDT)
From: Johannes Schindelin <[hidden email]>
To: "O'Malley, Donald" <[hidden email]>
Cc: [hidden email]
Subject: Re: [CONFOCALMICROSCOPY] I FOUND the little SQUARES

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Donald,

On Tue, 17 Apr 2012, O'Malley, Donald wrote:

> I get the point that the sharp boundaries we see between adjacent pixels
> are physically impossible (in this case, for sure), but seeing the
> actual recorded data, presented as recorded, was useful at least to me.
> Of course, the colors in the images are a figment of our artistic
> imaginations!

It is okay to visualise your data in whatever form gives you the most
insight. For example, I am a total fan of lookup tables (e.g. the Glasbey
LUT to catch, uhm, "creative" use of Photoshop). We are never looking at
the original data again, after all, when we look at digital images. The
important point is -- as you did -- to keep in mind what those pixel
values actually stand for (and to make it easy for others to keep that in
mind, too).

Ciao,
Johannes

[Mark Cannell]

*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
*****

Hi Johannes

I can certainly appreciate the problem of thresholding to decide which pixels are inside and which are outside an object and the problem of of edge aliasing (been there done that) , but your generalization about square pixels is not correct. If it was, you have just put the entire NASA JPL land sensing team out of business. I suggest you might like to look at some of their publications? Voxels are only cubes if the x,y,z ratios are equal (after correcting for refractive index differences of course).  Perhaps you would like to expand on your perceived problem?

Cheers

On 17/04/2012, at 5:39 PM, Johannes Schindelin wrote:

> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi list,
>
> Sorry, was sending from the wrong address, so the list rejected it.
> Therefore I am resending.
>
> But this gives me the opportunity to state my point more clearly:
>
> I do not take exception with displaying pixels as squares if that
> helps analyze the data.
>
> I do take exception with people pretending that pixels _are_
> square. They are not.
>
> I saw my share of bad analyses that were based on that false notion that
> pixels have an exact square-shaped extent (or for that matter, voxels that
> have an exact cuboid-shaped extent). For example, volume estimates of
> reconstructed objects of interest were wildly off because of that (and
> were reported without any indication of the expected precision), it was
> not even funny.
>
> Ciao,
> Johannes
>
> ---------- Forwarded message ----------
> Date: Tue, 17 Apr 2012 10:21:04 -0500 (CDT)
> From: Johannes Schindelin <[hidden email]>
> To: "O'Malley, Donald" <[hidden email]>
> Cc: [hidden email]
> Subject: Re: [CONFOCALMICROSCOPY] I FOUND the little SQUARES
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> *****
>
> Hi Donald,
>
> On Tue, 17 Apr 2012, O'Malley, Donald wrote:
>
>> I get the point that the sharp boundaries we see between adjacent pixels
>> are physically impossible (in this case, for sure), but seeing the
>> actual recorded data, presented as recorded, was useful at least to me.
>> Of course, the colors in the images are a figment of our artistic
>> imaginations!
>
> It is okay to visualise your data in whatever form gives you the most
> insight. For example, I am a total fan of lookup tables (e.g. the Glasbey
> LUT to catch, uhm, "creative" use of Photoshop). We are never looking at
> the original data again, after all, when we look at digital images. The
> important point is -- as you did -- to keep in mind what those pixel
> values actually stand for (and to make it easy for others to keep that in
> mind, too).
>
> Ciao,
> Johannes