OMERO.server image rendering

A major requirement for any image data application is the ability to display images. In most applications, this is achieved by reading pixel data from a filesystem and then mapping the pixel data to the 256 grey level available on most computer display monitors. It is common in some experiments to record and display multiple channels at once. Typically three, four, or even five separate images must be mapped, and then presented as a color image for painting on a monitor. Because these operations can require many thousands of operations and must be displayed rapidly to support the display of time-lapse movies, most image display software applications use a high-speed graphics CPU and dedicated hardware for image rendering and display. This requirement limits the deployment of these applications to high-powered workstations.

OMERO.server includes an image server, a software application that delivers rendered images to a client. This ensures that client applications can display image data. The OMERO Rendering Engine (OMERO-RE) has been designed to minimize the amount of data transferred to the client and thus removes the requirement for a specific graphics CPU, allowing high-performance image viewing on standard laptop computers. The OMERO-RE achieves this by limiting data transfer times by being close to the data, using highly efficient network transfer protocols, utilizing modern multi-processor and multi-core machines to provide the data to clients in a format that is as efficient to display as possible. OMERO-RE is multi-threaded and can use multi-core servers to simultaneously render individual channels before assembly into a final color image ready for transfer to the client. The use of the RE is not mandatory. If a client needs to have the full pixel data, it can. This OriginalPixels facility is used for client-side analysis, like that performed in the OMERO.insight measurement tool.

Transfer of image data even after rendering can limit performance, especially when accessing data remotely on connections with limited bandwidth (e.g. domestic ADSL). Therefore the OMERO-RE contains a compression service with an API that allows a client adjustable compression providing minimal image artefacts and a 20-fold range of data size to the client.

The OMERO Rendering Engine is accessed by OMERO client applications written in Java, C++, or Python via a binary protocol (ICE) provided by ZeroC.