Image Analysis and OMERO
While OMERO is not an image analysis software itself, OMERO supports to perform image analysis on the images stored on OMERO’s file system through programmatic connections with popular image analysis software or through user-friendly graphical user interfaces. That means, for most data, it is not necessary to download a copy of the data from OMERO to a local disk before image analysis can be performed. Rather, image analysis software can retrieve images directly from OMERO, even in batch, alleviating the workload significantly. Image analysis results, including generated regions of interest (ROIs) can be stored back in OMERO keeping analysis outcomes and original images closely linked.
Accessing OMERO from image analysis software
OMERO-hosted data can be accessed with a number of image analysis software (see image for a selection, taken from Bortolomeazzi & Boissonnet), each of which has its own strengths for image analysis tasks. The most popular open-source software for image analysis is ImageJ or Fiji (Schindelin et al., 2012; Jamaili et al., 2022; Schmidt & Hanne et al., 2022). When using Fiji, the connection to OMERO is enabled through a Fiji plugin, that can be downloaded from the OME website. On top, there is the OMERO-Fiji processing extension developed by Pouchin et al., which allows to even run macro-assisted image analysis with Fiji on OMERO hosted data in batch mode.
Regardless of the specific image analysis software, the general steps include:
1. Connect from the image analysis software to OMERO
2. Identify the image(s) to perform analysis on in OMERO
3. Run the image analysis in the software
4. Save the results back to OMERO or externally
5. Close the connection
Data hosted in OMERO is loaded over the network to be displayed in the image analysis software on a scientist’s computer. That means network speed and file sizes limit how well images can be loaded this way. A typical confocal microscopy image of ~20 MB usually works seamlessly.
The advantage of image analysis in combination with OMERO
Access to a specific image on a scientist’s computer is based on the image’s file path on the computer’s storage device. In point-and-click workflows, the user opens images manually from their file locations using the software’s “Open” function. In batch mode (i.e., processing and analyzing multiple images sequentially and automatically), each file is typically identified by its full file path (i.e., “hard-coded” in the script or pipeline). If images are later moved to a different folder or the folder hierarchy is reorganized, their file paths change. This breaks the hard-coded references, potentially causing the image analysis pipeline to fail to locate the files.
In contrast, OMERO uses a database. Each image has a unique, persistent ID that does not depend on its position in any organizational structure. Images grouped into a dataset can be identified through the dataset’s ID, and structured metadata annotations in OMERO, such as tags and key-value pairs, allow flexible data organization without relying on deep folder hierarchies. This means that images for an analysis process can also be identified via their associated metadata, independent of file system paths. Moreover, by leveraging OMERO’s central storage, multiple users can simultaneously perform image analysis on the same data without needing to maintain local copies.
From here, let’s continue with describing a discrete example of using Macro-based image analysis workflows in Fiji in combination with OMERO.
Back to: Bioimage Analysis
