[3dem] Instrumental Resolution versus Results Resolution

[Marin van Heel]

Dear All,

 From the lectures and discussions at the recent 3DEM GRC (Les 
Diablerets, 2017) I noticed there are still huge misconceptions – even 
among distinguished professors in Physics and  Biology - on what 
“resolution” means. So allow me to go over some basic principles:

1)The /instrumental resolution/ of an imaging device is given by the 
physical properties of the microscope, telescope, of whatever your 
favorite 1D-, 2D-, 3D-, 4D-imaging device is.The classical case would be 
that of a light microscope where the numerical aperture (NA) of the 
objective lens (https://en.wikipedia.org/wiki/Numerical_aperture) 
determines the “instrumental resolution” of the microscope 
(https://en.wikipedia.org/wiki/Angular_resolution). In the case of a 
diffraction-limited telescope it is the diameter of the main lens that 
determines the instrumental resolution. In the old days of Electron 
Microscopy one would often see the first zero of the CTF being used to 
define its instrumental resolution.

2) The /resolution/ achieved in the /results,///from images produced by 
our imaging device, is a very different issue! Suppose, for example, you 
forget to switch on the illumination of your light microscope! What good 
will then the high-resolution (high NA) properties of your expensive 
instrument do you? If, on the other hand, you do switch on the 
illumination but only use a very low dose of say 10,000 photons to 
generate an image, that image will be very noisy. How much better will 
the image of your object be if the image is instead created accumulating 
10,000,000,000 photons? The underlying question is: how do I define a 
results-related quality metric that reflects the image information I 
have collected in an experiment rather than what a specific instrument 
can potentially collect?

The basic idea here is to take TWO images of the same object rather than 
just ONE. Both images will contain the same signal (the object of your 
affection) but a different realization of the random noise so we can 
then compare the two images to each other in Fourier space using the FRC 
(Fourier Ring Correlation). This suggestion first emerged in 
single-particle EM in the early 1980s 
(https://en.wikipedia.org/wiki/Fourier_shell_correlation) 
<https://en.wikipedia.org/wiki/Fourier_shell_correlation%29>.

Strangely enough it took decades for the rest of the imaging scientists 
to realize what they were missing. Only very recently “everybody” 
suddenly started using the results-oriented FRC and FSC metrics in many 
other imaging fields, including X-ray microscopy, X-ray crystallography, 
light-microscopy, X-ray tomography, scanning microscopy, astronomical 
imaging, etc.Instead of claiming “super resolution” by showing some nice 
images from a given microscope, one can now objectively underpin that 
claim through the experimental FRC/FSC curve. I never understood why it 
took everybody so long to adapt to this straightforward gold-standard 
metric.

Take home lesson: there is the “INSTRUMENTAL RESOLUTION” that the 
imaging instrument has intrinsically, whether you actually use it or 
just leave it in the cupboard, and there is the statistically 
significant “RESULTS RESOLUTION” which reflects the quality of the final 
results achieved within a given data collection experiment. TWO very 
different concepts …!

My TWO cents,

Marin

-- 
==============================================================

     Prof Dr Ir Marin van Heel

     Research Professor
     Laboratório Nacional de Nanotecnologia - LNNano
     CNPEM/LNNano, Campinas, Brazil
     Brazilian mobile phone  +55-19-981809332
                            (041-19-981809332 TIM)
     Skype:  Marin.van.Heel
     email:  marin.vanheel(A_T)gmail.com
     and:    mvh.office(A_T)gmail.com

----------------------------------------------

     Emeritus Professor of Cryo-EM Data Processing
     Leiden University

----------------------------------------------

     Emeritus Professor of Structural Biology
     Imperial College London

--------------------------------------------------------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20170624/d0bff8a9/attachment.html>