*****
To join, leave or search the confocal microscopy listserv, go to:
http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
Post images on http://www.imgur.com and include the link in your posting.
*****

 This is great! I didn't know there were good, fast detectors with just a
few pixels. I wonder what they used, and if they're for sale? If the
detector gives pixels per second and noise similar to a decent camera, this
could be a really nice scope.

 I'm also curious about their deconvolution algorithm. It sounds like it
might be similar to Maria's approach (
http://dx.doi.org/10.1002/cphc.201300831 , figure 3), or maybe just
separate deconvolution and averaging of each individual "pinhole"? If you
want to do this fast and well, you almost certainly need tons of
parallelization.

Related:
 There seems to be disagreement if (ISM/msim/instant sim/rescan
confocal/OPRA/etc) improves axial resolution. This is actually a fun
question, and I think the answer is, surprisingly, "it depends on the
detailed shape of your excitation and emission PSFs".

 It's clear the (sheppard/meuller/york/de luca/roth) trick of scaling the
image of the laser spot relative to the final image doesn't change
(physical) sectioning*: all the light that passes the (macro)pinhole hits
the detector, whether or not you scale the image of the laser spot. The
interesting question is: does the scaling trick cause the image of a bead
to diverge any faster as you defocus in Z? If it does, your axial
resolution is better, even if your sectioning is the same**.

 Suppose your emission and excitation PSF both look like Gaussian beams
(even though they probably don't). If so, Enderlein's analysis holds,
whether or not the beads are in focus, and the scaling trick gives the same
lateral resolution improvement regardless of defocus. If out-of-focus beads
are sharpened just as much as in-focus beads, the bead images don't diverge
any faster or slower compared to normal confocal - meaning no improvement
in axial resolution.

 However, suppose your emission or excitation PSFs only resemble Gaussian
spots when they're in focus. Then, Enderlein's analysis fails when the
beads are not in focus, and the scaling trick won't work as well for
out-of-focus beads. This means the bead images diverge faster, compared to
normal confocal - and your axial resolution is improved!

 Anyhow, happy to see the field moving along, eager to hear more details
about Zeiss's scope as they become available.

* By "sectioning", I mean the ability to resolve two closely spaced thin
uniform fluorescent z-planes. By "physical sectioning", I mean rejecting
the noise from out-of-focus light like a confocal, not subtracting off the
signal and leaving the noise like an apotome.
** By "axial resolution", I mean, the ability to resolve two closely spaced
fluorescent beads in Z.


On Sat, Jul 26, 2014 at 8:57 AM, Guy Cox <[log in to unmask]> wrote:

> All true, but I'd like to give credit also to Carol Cogswell, when she was
> working with Colin Sheppard in Sydney in the 90s.  She was interested in
> transmission confocal, where the specimen's refractive properties tend to
> deflect the spot.  Her idea was to use an array to follow the spot as it
> wandered.
>
>                                                 Guy
>
> Guy Cox, Honorary Associate Professor
> School of Medical Sciences
>
> Australian Centre for Microscopy and Microanalysis,
> Madsen, F09, University of Sydney, NSW 2006
>
> -----Original Message-----
> From: Confocal Microscopy List [mailto:[log in to unmask]]
> On Behalf Of Shalin Mehta
> Sent: Saturday, 26 July 2014 1:00 PM
> To: [log in to unmask]
> Subject: Re: Johannes Amon posted new Zeiss white paper on air yscanning
> at ...
>
> *****
> To join, leave or search the confocal microscopy listserv, go to:
> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> Post images on http://www.imgur.com and include the link in your posting.
> *****
>
> It is interesting to see this idea go commercial. We computed transverse
> and axial-resolution of this scheme (which we call pixel
> reassignment):
> https://www.researchgate.net/publication/254260061_Superresolution_by_image_scanning_microscopy_using_pixel_reassignment?ev=prf_pub
> We found that (Fig. 5 of above paper) if the detector array is close to
> the size of the airy disk, you actually don't lose too much sectioning, but
> gain significant SNR, as also noted in the white paper.
>
> The original idea of pixel reassignment was proposed by Colin Sheppard in
> 1988:
>
> https://www.researchgate.net/publication/235994020_Super-resolution_in_confocal_imaging
>
> Shalin
> On Fri, Jul 25, 2014 at 8:33 AM, John Oreopoulos <
> [log in to unmask]> wrote:
> > *****
> > To join, leave or search the confocal microscopy listserv, go to:
> > http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> > Post images on http://www.imgur.com and include the link in your
> posting.
> > *****
> >
> > Zdenek,
> >
> > I think you're right. When I read this white paper again closely, it
> actually sounds more like what was recently branded "re-scan confocal
> microscopy":
> >
> > De Luca, G. M. R., Breedijk, R. M. P., Brandt, R. A. J., Zeelenberg, C.
> H. C., de Jong, B. E., Timmermans, W., et al. (2013). Re-scan confocal
> microscopy: Scanning twice for better resolution.Biomed. Opt. Express, 4,
> 2644-2656.
> >
> > A lot of the same references are in this article, and this also cites
> Enderlein and York.
> >
> > Cheers,
> >
> > John Oreopoulos
> >
> >
> > On 2014-07-25, at 4:09 AM, Zdenek Svindrych wrote:
> >
> >> *****
> >> To join, leave or search the confocal microscopy listserv, go to:
> >> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> >> Post images on http://www.imgur.com and include the link in your
> posting.
> >> *****
> >>
> >>
> >> Very interesting.
> >>
> >> It's most likely based on the Image Scanning Microscopy (ref 4 of the
> >> white paper; it's a physics journal, so the style is a bit
> >> different), it allows up to 2x resolution improvement (as all the
> >> other linear structured illumination methods), but with the
> >> Enderlain's image processing you actually sacrifice axial resolution
> >> if you want high SNR (corresponds to bigger pinhole). Unfortunately
> >> they don't provide much details in the white paper about the new
> >> image processing approach that gives also axial superresolution...
> >>
> >> Moreover, I think this approach is not quite compatible with other
> >> benefits of modern confocals, such as spectral detection or lifetime
> imaging.
> >>
> >> Looking forward to seeing further advances in this direction.
> >>
> >> Zdenek Svindrych, Charles univ., Czech Rep.
> >>
> >>
> >>
> >> ---------- Původní zpráva ----------
> >> Od: John Oreopoulos <[log in to unmask]>
> >> Komu: [log in to unmask]
> >> Datum: 25. 7. 2014 4:30:32
> >> Předmět: Re: Johannes Amon posted new Zeiss white paper on airyscanning
> at .
> >> .
> >>
> >> "*****
> >> To join, leave or search the confocal microscopy listserv, go to:
> >> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> >> Post images on http://www.imgur.com and include the link in your
> posting.
> >> *****
> >>
> >> The white paper states:
> >>
> >> "An acentric, shifted pinhole detector produces an image of about the
> >> same resolution as a pinhole detector which is aligned to the optical
> >> axis, although smaller in amplitude and shifted by half the
> >> displacement. This insight has been the motivation for constructing
> >> an area detector for a confocal microscope. Such a detector should
> >> cover more than 1 AU and contain multiple sub-Airy detector elements.
> >> Detection efficiency will be significantly increased by reassigning
> >> the detected photons from the shifted detector elements to the
> >> central detection position and summing up the back shifted signal
> >> from all detector elements. [2] No light is rejected by a closed
> pinhole but instead collected by the off-axis detector elements.
> >> Therefore an increased signal level arises from the reassignment of
> >> photons to a smaller spatial region."
> >>
> >> It sure does sound a lot like Jim Pawley's array detector pinhole
> >> idea to me, which is outlined in quite clearly chapter 2 of the
> >> Handbook of Biological Confocal Microscopy (3d Edition):
> >>
> >> http://link.springer.com/chapter/10.1007/978-0-387-45524-2_2
> >>
> >> See Figure 2.15 in particular. There isn't much detail in the white
> >> paper to know what the differences are however, and there is no
> >> citation of Pawley's work on this. Here's the relevent passage from the
> chapter:
> >>
> >> "In addition to high QE, Si photon detectors have a variety of other
> >> practical advantages. As the sensitive element in such a detector is
> >> typically very small (5–30 mm on a side), selective use of only a few
> >> elements in a small, planar, 2D array could permit it to operate in
> >> the CLSM as a combination pinhole and detector. Figure 2.15 is a
> >> sketch of what such a detector might look like. After each pixel
> >> interval of the microscope, the charge pattern in the 5 x 5 sensor
> >> array at the top would be transferred to the read register and then
> >> the signal in all 25 pixels would be read out sequentially at about
> >> 35MHz. These 25 values could then be “decoded” in a number of
> >> possible ways, the most straightforward of which would be to provide
> >> three separate signals corresponding to the summed signals from the
> >> brown, orange, and red areas of the sensor array. In this way, it would
> be possible to collect signal simultaneously at three different pinhole
> sizes.
> >> With such a detector, pinhole alignment could be done elec-
> >> tronically simply by searching for the detector element producing the
> >> most signal from a planar specimen and misalignment could be detected
> >> on the fly by comparing, for example, summed output from the 5 pixels
> >> on the left with the
> >> 5 on the right (Pawley, 1996)."
> >>
> >> The other reference from 1996 has even more details about this idea:
> >>
> >> CCDiode: an optimal detector for laser confocal microscopes
> >> http://spie.org/Publications/Proceedings/Paper/10.1117/12.237490
> >>
> >> John Oreopoulos
> >>
> >>
> >>
> >> On 2014-07-24, at 8:57 PM, George McNamara wrote:
> >>
> >>> *****
> >>> To join, leave or search the confocal microscopy listserv, go to:
> >>> http://lists.umn.edu/cgi-bin/wa?A0=confocalmicroscopy
> >>> Post images on http://www.imgur.com and include the link in your
> posting.
> >>> *****
> >>>
> >>> Johannes Amon posted the URL to download a new Zeiss white paper on
> >> airyscanning
> >>>
> >>> http://www.zeiss.com/airyscan
> >>> <https://www.linkedin.com/redirect?url=http%
> >> 3A%2F%2Fwww%2Ezeiss%2Ecom%2Fairyscan&urlhash=fM6m&_t=tracking_anet>
> >>>
> >>> Johannes post is at
> >>>
> >>> https://www.linkedin.com/groupAnswers?viewQuestionAndAnswers=&discus
> >>> sionID
> >> =5898093815864004612&gid=837267
> >>>
> >>> and he welcomes comments.
> >>>
> >>>
> >>> Airyscanning: A Novel Approach to Confocal Imaging
> >>>
> >>> Johannes Amon
> >>> <https://www.linkedin.com/profile/view?id=86600201&goback=%2
> >> Egde_837267_member_5898093815864004612>Online Communications and
> >> Marketing
> >>>
> >>> It is with great pleasure that I am able to share with you a new
> >>> White
> >> Paper by ZEISS. Airyscanning is a new detection concept that uses an
> >> array detector to oversample each Airy disk in order to gain
> >> sensitivity, resolution and speed. I'm curious to hear what you think
> >> of it! Download the free White Paper here:http://www.zeiss.com/airyscan
> <https://www.linkedin.
> >> com/redirect?url=http%3A%2F%2Fwww%2Ezeiss%2Ecom%2Fairyscan&urlhash=fM
> >> 6m&_t=
> >> tracking_anet>
> >>>
> >>>
> >>> Vladimir Zhukarev and I have already commented at the linkedin
> >>> group. My
> >> comments are:
> >>>
> >>>
> >>>
> >>> Zeiss cites some but not all of the literature of using a (small)
> >>> area
> >> array to collect light that would usually be blocked by the pinhole.
> >> I'm thinking that Jim Pawley had this, but could be 'off by one'
> >> (Tony Wilson, Guy Cox, etc?).
> >>>
> >>> Too bad Zeiss and the other microscope companies have pathetically
> >>> slow
> >> spatial deconvolution AND on confocal microscopes deconvolution is a
> >> separate post-processing step the user has to undertake manually (I
> >> managed Leica confocals since 2000 and Zeiss LSM710 since 2009 ...
> >> just trained this week on a Leica SP8 confocal, LAS AF 3.x, haven't
> >> seen all the features, but deconvolution looks like still a separate
> manual step).
> >>>
> >>> Further, to get the maximum benefit from a confocal microscope, when
> >> acquiring multiple channels, should BOTH spatially deconvolve AND
> >> spectrally unmix. See the Hoppe et al 2008 Biphys J paper,
> >>>
> >>> http://www.ncbi.nlm.nih.gov/pubmed/18339754
> >>>
> >>> their 3DFSR was tested with FRET specimens but the (claimed) 10x
> >>> improved
> >> SNR should apply to any specimens.
> >>>
> >>> I also note that the Zeiss (claimed) 1.7x improvement could be
> >>> massively
> >> improved by acquiring 300 images (per plane) and doing 3B microscopy
> >> or SOFI
> >> - for 3B see
> >>>
> >>> http://www.coxphysics.com/3b/
> >>>
> >>> for introduction to 3B (and yes, could work in 3D). 3B is also very
> slow.
> >> Parallel processing is speeding up 3B, see
> >>> http://www.optnano.com/content/2/1/7/abstract
> >>>
> >>> For an entry point to SOFI see
> >>>
> >>> http://www.optnano.com/content/2/1/2/abstract
> >>>
> >>> As I've noted on the Confocal Listserv
> >>>
> >>> http://lists.umn.edu/cgi-bin/wa?A2=ind1407&L=CONFOCALMICROSCOPY&P=59
> >>> 20
> >>>
> >>> I am psyched what parallel processing is doing (ex. NVidia TITAN Z
> >>> GPU
> >> card, Intel Phi 7120 card, lots of Phi's, as at
> >> https://www.tacc.utexas.edu/ stampede/ ) and will get even better
> >> with 2015 Knights Landing CPU's (and NVidia, IBM, HP if its "The
> Machine" meets their goals, etc).
> >>>
> >>> George
> >>>
> >>>
> >>> www.linkedin.com/in/georgemcnamara/
> >>>
> >>> --
> >>>
> >>>
> >>>
> >>>
> >>> George McNamara, Ph.D.
> >>> Single Cells Analyst
> >>> L.J.N. Cooper Lab
> >>> University of Texas M.D. Anderson Cancer Center Houston, TX 77054
> >>> Tattletales http://works.bepress.com/gmcnamara/42"
>