Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
Thanks for the explanation, John. Are you sure no other image processing
was done other than bicubic resampling for the inset? There seems to be
something going on between panels B and C, before even looking at the inset.
1) The image in C looks less noisy than the image in B. This is
counterintuitive. The ratio of two images usually looks noisier than the
overlay of the same two images, although it certainly depends on how you map
the pixel values to the color palettes you choose. However, since you
choose such a wide range of colors for panel C compared to B, I would expect
the noise to be more pronounced than in B.
2) The structures in C are dilated with respect to the visible structures in
B. Many of the variations you see in pH on the edges of the neurons are
thus in areas where there is very little if any signal.
Bicubic resampling, the way I understand it, merely interpolates between the
pixels, so it shouldn't change the gradients between the original pixels.
However, I can see that it might dilate objects if the threshholds are not
chosen carefully.
I appreciate any efforts by authors to explain the exact image processing
that goes on with their data.
Cheers,
Jeff
> -----Original Message-----
> From: Carson,John H. [mailto:[log in to unmask]]
> Sent: Monday, December 13, 2004 11:30 AM
> To: [log in to unmask]
> Subject: Re: Better than possible confocal images?
>
>
> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=3Dconfocal
>
> I have followed this confocal listserver for many
> years and = benefited from the collective wisdom of the
> members on numerous = occasions. However I never expected to
> find one of my papers as a = thread topic.
> The cover image, to which Dr. Schweining refers in
> this thread, = was provided in response to a request from the
> journal for a figure = illustrating the findings in our paper
> on pH microdomains in = oligodendrocytes. The image was
> derived by scaling a region from Figure = 1C using bicubic
> resampling. This procedure smooths the boundaries = between
> individual pixels in the original image making the =
> intracellular pH gradients appear more smoothly continuous
> than in the =
> original image. I agree that this does produce an image that is =
> "better than possible by confocal imaging". Our rationale
> for using = this procedure to generate the cover image was to
> provide an = illustration of inracellular pH gradients which
> we believe are likely to = be continuous rather than
> pixellated. We should have provided a more = detailed
> description for how the cover image and the inset to Fig 1C =
> were generated.
>
> John Carson=20
>
> > ----------
> > From: Confocal Microscopy List on behalf of
> Christof Schwiening
> > Reply To: Confocal Microscopy List
> > Sent: Monday, December 6, 2004 11:48 AM
> > To: [log in to unmask]
> > Subject: Better than possible confocal images?
> >=20
> > Search the CONFOCAL archive at
> >http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=3Dconfocal
> >=20
> > Dear All,
> >=20
> > I have come across a paper that claims to show regional pH =
> measurements
> > (confocal fluorescence, ratiometric, of a pH indicator) with =
> sub-micron
> > resolution and no apparent noise from single images of live
> cells. The
> > images show pH gradients of about 1 pH unit over distances
> less than 1
> > =
> um
> > within the cytosol. The paper is Ro & Carson 2004 JBC =
> 279(35):37115-23. The
> > pH gradients are shown on the cover of the 27th August
> edition of JBC.
> >=20 The PubMed link is:
> > http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?
> > cmd=3DRetrieve&db=3Dpubmed&dopt=3DAbstract&list_uids=3D15192115
> >=20
> > If you look at the inset to Fig 1C it shows a region 5um
> long with pH =
> 7.5
> > regions sitting right next to pH 6.5 regions.
> >=20
> > Are these sorts of images, with very high resolution and no noise =
> possible?
> >=20
> > Does anyone feel able to comment?
> >=20
> > Greetings,
> > Christof
> >=20
> >=20
>
|
