[3dem] EPU and serialEM

[Vladan Lucic]

Hi David,

Were the images recorded with different image shifts? If that was the 
case, the stage coordinates alone would not be sufficient to predict the 
pixel positions of the data acquisition images on the grid square image. 
It could also be a combination of inaccurate pixel size and defocusing, 
image rotation, ... .

We earlier reported two methods to establish a correlation between 
low-mag and high-mag images (these roughly correspond to your grid 
square and data acquisition images, respectively), see Coordinate 
transformation based cryo-correlative methods for electron tomography 
and focused ion beam milling, Fukuda et al 2014. Feel free to contact me 
if you need more info.

Best,
Vladan

On 06/09/2023 18:29, Morgan, David Gene wrote:
> Hi,
>
> I'm cross-posting this both to the 3dem and the serialEM lists in the 
> hope that someone may be able to help me.  It concerns data from our 
> Talos Arctica acquired with EPU that I would like to be able to make 
> serialEM aware of.
>
> For reasons I won't bore you with, I am trying to take the stage 
> positions acquired during an EPU data acquisition run and map them 
> onto the image of a grid square also acquired during that same run.  
> In other words, from the XML files that EPU produces, I can extract 
> the stage coordinates of both the data acquisitions and the grid 
> square image, and I want to "unify" them at the level of the grid 
> square image.
>
> With a bit of math, I can convert the stage coordinates of the 
> acquisitions into pixel coordinates for the grid square image. I'm 
> initially overlaying these coordinates onto the grid square image 
> using IMOD, after which I will move to making serialEM aware of this 
> information. When I map locations onto the grid square image, the 
> resulting locations don't make much sense... For example, some of the 
> acquisitions are deep into grid bars and none of them align with holes 
> in the holey carbon film. Something is clearly very wrong (and I don't 
> think it's the relatively simple math).
>
> It could be a simple as an incorrect pixel size for the grid square 
> image (but it would require that the pixel size be incredibly wrong), 
> or it may be something going on inside EPU that I don't understand 
> (like the application of some sort of transformation between the EPU 
> imaging states - I'm thinking of the steps to align the various images).
>
> Has anyone attempted anything like this, or does anyone have a clue 
> about why the results seem so "off"?
>
> Thanks in advance for any insights.
>
> -- 
>     politics is more difficult than physics.
>                                              A. Einstein
>
>             David Gene Morgan
>         Electron Microscopy Center
>              047E Simon Hall
>              IU Bloomington
>           812 856 1457 (office)
>           812 856 3221 (3200)
> https://urldefense.com/v3/__http://iubemcenter.indiana.edu__;!!Mih3wA!B99efhbjH-KcQtsm8-yPK--pB67agdVu__TMVGwPnZxDApgX6tkWfhXPVDAnpsMZiBG2d3fxYUapskxuqdD4p4s$  
> <https://urldefense.com/v3/__http://iubemcenter.indiana.edu__;!!Mih3wA!DxfIH-v-lBNKsFnj4GP6Ttx7cUNj5MPm8GTbHzChW3LUYYypiZQvnaWBJJRJslcHlN-fxG6SoZTIU4Hc3fXkoFI$>
>
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20230907/c8d19733/attachment.html>