FIJI bug: imported MRI slice skewed due to wrong automatic dimension

fiji

#1

When I use Import/Image (the only option that works for the image) to open an MRI slice from a Fuji system, the image appears but is skewed badly, probably due to the image import reading the number of columns incorrectly. I can send an example file that demonstrates the problem, but I see no way to include that file in this message because, as is standard, it has no extension and your upload does not allow a file with no extension.


#2

Hi @TimM

have you tried the import options on this website?

https://imagej.net/DICOM

If they don’t work would you mind sharing an example data set with us?
I hope that helps.

Cheers,
Robert


#3

Robert - Thank you for taking the time to answer my query!

The BioFormats import does not work at all for these images. In fact, it appears that none of the import options other than ‘Image’ reads the file. My guess is that the ‘Image’ option would read it perfectly if it had the correct column count, as the image looks correct except for the skew.

The Tudor tools look promising, but the link for downloading them does not appear to lead to a download site.

I have attached a sample image that displays the problem. Thank you!

Timothy Masters

(Attachment 00000001 is missing)


#4

Robert - Your email system rejected my second attempt to attach a sample image file that demonstrates the problem. It says that new users are not allowed to send attachments. Do you have a personal email where I can send it? Thanks!

Oh… here’s another approach. I uploaded it to my cloud storage site. You can download it via this link: (I’m replacing the dots with ‘DOT’ in case your email disallows links.)

https://files DOT secureserver DOT net/0sa7eKzP41H26m

Tim


#5

Hi @TimM

have you tried this link?

http://santec.tudor.lu/project/dicom/download

Cheers,
Robert


#6

Robert - Thank you for taking the time to give me all that help!

At least for the moment I found a quicker and easier workaround instead of installing large new programs.

I just exported the skewed MRI slice (DICOMOBJ) from FIJI, saving it as a BMP. Then I wrote a trivial little program that goes into the BMP header and changes the 576 columns reported by FIJI to the actual 575 columns. That gives me a correct image file that is easily read into FIJI using the simple File / Open command.

If you feel it appropriate, perhaps you could forward to the person who wrote the import code the link I previously sent that allows the user to download the DICOMBOBJ file whose column count is being incorrectly obtained. This is a bug that he/she might want to address.

Thank you once again!

Tim