« Everything you need to know about image screening at Rockefeller University Press in 10 posts | Main | Everything you need to know about image screening at Rockefeller University Press in 10 posts »

October 12, 2015

Comments

Will_j_moore

What about microscope images, where "dpi" has no meaning? You simply acquire a 512 x 512 image, with no known print size. How do you ensure it is of sufficiently high resolution?

Melina Casadio

Hi Will,

Thank you for your comment! You are absolutely right that, for microscope data acquisition, "resolution" is indeed dependent on the hardware – one cannot set a resolution in dpi for these images on the scope, as you know.

However, a field captured at 512x512 for instance should still yield an image of appropriate resolution. If you open the newly acquired micrograph in an Adobe program (for instance) and save the file as a TIFF file (without compression or if needed using LZW), you will be able to check the size of the image. A 512x512 capture will yield an image of relatively small size as measured in inches, but it is acceptable for publication. You can of course edit the the properties of the image (size/resolution) and set them to the desired settings.
If the image is straight from the microscope and you follow these steps immediately without first altering the micrograph using other software or compressing it to reduce file size, it’s unlikely such an image would be of insufficient resolution – even if at the maximum zoom and/or at the “resolution limit” of the microscope. (As you know, if you capture a large field and need to zoom in, your available resolution diminishes, so if you want both a broad view and a close-up of one section, we suggest making two captures.) We found that most micrographs of "low resolution" are of poor quality because they were saved/stored as JPEGs and/or heavily compressed, and these issues can be avoided by following these steps. We’re of course happy to discuss this question further on a case by case basis.

Thank you again for your interest in the column and for the question!

Will_j_moore

Certainly a 512 x 512 image is likely to be high enough resolution if you want to show the whole image in the figure, since at 300 dpi this will be 1.7 inches on the page. Or at 1 inch on the page this is 512 dpi. However, if you wish to crop and zoom a portion of the image, to show a feature at higher magnification (E.g. a centromere at 8 times the size of the parent image), then this could produce an 'inset' as low as 64 dpi (512 dpi / 8). In this case, you will be forced to increase it's resolution and if you want to maintain a non-pixelated appearance you will also have to re-sample the inset (add dots). JCB rightly instructs users not to re-sample images, but unfortunately this is not enforced in JCB or any other journals, since a non-pixelated image is considered preferable to doing the "right thing".
I would prefer it if the instructions better matched what authors are actually doing to get the 'nice' figures required for publication, and include full details of the figure preparation steps.

I should say that I have a particular interest in this subject since I am the developer of a figure preparation tool that does just thishttp://figure.openmicroscopy.org/. I'm trying to give users the tools they need (including re-sampling) without breaking the guidelines stipulated by journals.

Melina Casadio

Thank you for your reply Will! You are right that we do not go into 100% of all the possible cases here -- this column is aimed at trouble-shooting our most common issues and only provides general tips without going into the detail of all possible figure issues. (Of note, there are a number of reasons for which we are not too strict about zoomed-in/high-magnification images during the figure-screening process, but that's off topic here.)
However, we are working on revamping our Instructions for Authors - so we'll take your suggestion under advisement! And please check back soon: we will discuss appropriate figure-making software soon and we happily recommend OMERO.figures among other options. Thank you again for checking out this series!

The comments to this entry are closed.