[3dem] -|EXT|- Re: Workflow for TEM volumes containing noise in Amira?

[HUGHES Louise]

Hi Charlotte,

I would also recommend IMOD’s filters. I found the median filter to be very useful in reducing noise and enhancing contrast in tomograms. As Pranav has indicated, you can try different settings to evaluate of them is best and you apply them using the clip function.

Amira also has some filters such as a gaussian filter that you can apply to the data.

Good luck!

Louise

From: 3dem <3dem-bounces at ncmir.ucsd.edu> On Behalf Of Pranav Shah
Sent: Tuesday, April 28, 2020 12:38 PM
To: Charlotte Hamngren <charlotte.hamngren at gmail.com>
Cc: 3dem at ncmir.ucsd.edu
Subject: -|EXT|- Re: [3dem] Workflow for TEM volumes containing noise in Amira?


Hi Charlotte,

In my (quite limited) experience, i have had some degree of success with:

1. Aggressively low-pass and smoothening the tomograms using imod’s clip filter utility. In order to work out the right values, open up the tomogram in 3dmod>image>process> filter. Tinker with the sliders until you are satisfied and note the lp value and the roll-off. Pass those values onto clip filter -l <lp cutoff>,<roll-off> inputTomo.rec outputTomo.rec

2. If you are comfortable with MATLAB, you can git clone dimitry tegunov’s tom and tom_deconv packages and use his deconvolution script. He has helpful note on how to use the script and is easy to follow.

If you are feeling up to it, install CryoCARE from florian jug’s lab to use deep-learning for denoising tomograms. I have not tried it myself, but the results look impressive and may aid in your segmentation routine.

Hope this helps,
Pranav

On Tue, 28 Apr 2020 at 11:54, Charlotte Hamngren <charlotte.hamngren at gmail.com<mailto:charlotte.hamngren at gmail.com>> wrote :
Dear 3DEM community,

I hope you are well!

I am using Inspect3D together with Amira for the 3D visualisation of membrane structures inside the a cell: ER and Golgi, etc. My raw data is a TEM tilt series (room temperature) acquired with the FEI Tomography software. The membranes in the reconstructed volume (SIRT 20 iterations) are clearly visible by eye, but the reconstructed volume is quite noisy since the sample cannot be heavily contrasted. I would like to automate the visualisation process as much as possible before turning to manual segmentation as a last resort. The Volume rendering process in Amira suffers a lot from the noise, so I am looking for advice. Is it a waste of time to try to de-noise the volume?

Is anyone else doing something similar? Can you suggest a usable workflow for working with TEM volumes containing noise in Amira (a publication or your own experience perhaps)?

Stay safe,
Charlotte
_______________________________________________
3dem mailing list
3dem at ncmir.ucsd.edu<mailto:3dem at ncmir.ucsd.edu>
https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
--
Best,
Pranav
--
Pranav Shah
Postdoctoral Research Fellow.

Division of Structural Biology,
Wellcome Trust Centre for Human Genetics,
University of Oxford,
Roosevelt Drive, Oxford OX3 7BN,
UK
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20200428/7bc45984/attachment-0001.html>