[ome-devel] BIOFORMATS bug report from a user: GATAN dm3 not read correctly

Tue Sep 9 17:46:51 BST 2014

Hello,

I uploaded the GATAN dm3 file containing the image with a negative value to a pixel set, and found this to be the reference to the image:  QA Bug | 9526 (new).

Let me take the chance and add some further, important observations I did today:

I found , that going in FiJi the following way, then the data import works correctly for my dm3 with a negative value stored for a pixel:

File à Import à Bio-Formats

After selecting the file of interest, a dialog window  "Bio-Formats Import Options" appears, and having answered the default settings the import does work correctly! This Options dialog appears always again if importing a dm3 file.

But, if only importing in FiJi data in the following way, then the bug appears the same as it appears using the MATLAB front end for the Bio-Formats import function 'bfopen':

File à Open...

does not work correctly.

Apparently, the FiJi "File à Open..." path uses some identical step as the MATLAB /Bio-Formats command 'bfopen' does, while the "File à Import à Bio-Formats" path does something different which works flawless.

I for sure downloaded and installed and use this MATLAB toolbox for using Bio-Formats in MATLAB and believe to have the older LOCI things deleted:

http://downloads.openmicroscopy.org/bio-formats/5.0.4/artifacts/bfmatlab.zip

I further believe to remember correctly that I downloaded and installed FiJi, and somehow remember that there was something about ImageJ2 support, which I for sure didn´t reject, whatever it was. If there has been any choice to get FiJi in different versions, then I believe I decided for the FiJi with version ImageJ2 in use. I remember quite well that I decided to not install ImageJ and LOCI plugin anymore but to install for convenience the all-inclusive package FiJi, because I read that then the LOCI bioformats would be integrated, and FiJi update mechanism would also update that bioformats part then.

I am now confused to have seen somewhere on the screen (not sure if it actually was in FiJI or MATLAB), that some SCIFIO is somewhere in use or has been mentioned in a message when opening / importing a dm3 file. I am further confused because I thought to have read somewhere that if the LOCI plugin would be installed that then ImageJ would automatically  use it in the standard "File à Open..." path if needed. I am completely confused if I am actually using now the LOCI plugin in ImageJ, or the OME Bio-Formats import filter, or some SciFio thing or whatever, but this should not be the topic here in my bug report.

In summary: for sure the MATLAB 'bfopen' command does not produce the correct result, and the FiJi "File à Open..." menu neither, but the FiJi "File à Import à Bio-Formats" menu apparently works correctly.

Best greetings from the EM-Labs of CIC biomaGUNE,

Marco Möller

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

Marco Möller

Platform Manager - Electron Microscopy

CIC biomaGUNE

Parque Tecnológico de San Sebastián

Paseo Miramón, 182. Ed. Empresarial C

20009 San Sebastián (Guipúzcoa)

SPAIN

Tel:  +34 943 00 53 25

mmoller at cicbiomagune.es <mailto:mmoller at cicbiomagune.es> 

From: Balaji Ramalingam [mailto:b.ramalingam at dundee.ac.uk] 
Sent: Tuesday, September 09, 2014 2:54 PM
To: Marco Moller; ome-devel at lists.openmicroscopy.org.uk
Subject: Re: [ome-devel] BIOFORMATS bug report from a user: GATAN dm3 not read correctly

Hi Marco,

Thank you for your detailed summary.

As suggested, it would be great if you could upload the file to the following URL,

https://www.openmicroscopy.org/qa2/qa/upload/

We will then create a ticket, post investigation and CC you on the same.

Thanks again,

Best,

Balaji

__________________

Mr Balaji Ramalingam

Software Developer

OME Team

College of Life Sciences

University of Dundee

From: Marco Moller <mmoller at cicbiomagune.es>
Date: Tuesday, 9 September 2014 08:22
To: "ome-devel at lists.openmicroscopy.org.uk" <ome-devel at lists.openmicroscopy.org.uk>
Subject: [ome-devel] BIOFORMATS bug report from a user: GATAN dm3 not read correctly

Hello,

I found a bug in the current BIOFORMATS import filter for GATAN *.dm3 (Electron Microscopy) files:

If the original GATAN software stored for a pixel in the image a negative value, then the BIOPFRORMATS import filter does not read this out correctly, but changes the negative value to an extreme high positive value.

I am working with MATLAB and have not much experience with ImageJ, but at least could confirm that this does not only happen with the MATLAB front end 'bfopen' of the BIOFORMATS import filter, but also happens when importing such image containing a negative value to ImageJ by the BIOFORMATS import filter used there.

I work on WinXP 32-bit or Win7 64-bit, have FiJi, MATLAB and BIOFORMATS software all in the most up to date version (as of 8 September 2014) installed.

Please have a look on the following little log on querying the max and min pixel value found in the image by different software:

Gatan Digital Micrograph V1.85 (32-bit):

GATAN_IMAGE: Minimum -93 at (1920,612)

GATAN_IMAGE: Maximum 1320 at (2035,135)

Gatan Microscopy Suite Digital Micrograph V2.02.800.0 (32-bit):

GATAN_IMAGE: Minimum -93 at (1920,612)

GATAN_IMAGE: Maximum 1320 at (2035,135)

MATLAB (64-bit), using BIOFORMATS command 'bfopen'

max = 4294967203 at (613,1921)

min = 45 at (1722,407)

after changing manually the (in MATLAB max,  in GATAN negative) value to i.e. 1000 using MATLAB command GATAN_IMAGE(613,1921) = 1000, applying 'bfopen' in MATLAB shows:

max: 1320 at (136,2036)

min: 45 at(1722,407)                       (which is correct, this I confirmed to before have been the second smallest value in the file)

Checking if this is a 'bfopen' behavior or a general BIOFROMATS behavior:

FiJi (64-bit), using the BIOFORMATS import option:

min:       45

max:      4,29E+09

I would be willing to share the GATAN image file of interest, but prefered not to attach it to the e-mail to the mailing list. If anyone would have interest in the image, please contact me and I can send it then to the interested person(s).

As I am meanwhile using the really excellent dm3 import filter written for MATLAB by Andreas Korinek (http://www.mathworks.de/matlabcentral/fileexchange/45933-gatan-digital-micrograph-file-reader), which I recommend you to have a look at in order to find much information about the dm3 file structure (much more data is extracted from dm3 than by the present BIOFORMATS filter), my e-mail to the developer list does not ask for your explicit answer. Instead it is cordially meant to just support the BIOFORMATS project by the given information and bug report.

Best greetings from the EM-Labs of CIC biomaGUNE,

Marco Möller

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

Marco Möller

Platform Manager - Electron Microscopy

CIC biomaGUNE

Parque Tecnológico de San Sebastián

Paseo Miramón, 182. Ed. Empresarial C

20009 San Sebastián (Guipúzcoa)

SPAIN

Tel:  +34 943 00 53 25

mmoller at cicbiomagune.es <mailto:mmoller at cicbiomagune.es> 

This e-mail is from CIC biomaGUNE. The e-mail and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. Any unauthorised dissemination or copying of this e-mail or its attachments, and any use or disclosure of any information contained in them, is strictly prohibited and may be illegal. If you have received this e-mail in error, please notify or telephone + 34 943 00 53 00 and delete it from your system.

The University of Dundee is a registered Scottish Charity, No: SC015096 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-devel/attachments/20140909/5d9466fd/attachment-0001.html>