OME-TIFF Overview and Rationale

Terminology
Image: a single (2D) image plane
Dataset: a collection of image planes composing a multidimensional structure (also called a multidimensional image or OME image)
OME dataset: a collection of related OME images
OME project: a collection of related OME datasets
Pixels: image samples acquired from a microscope or other source
Metadata: auxiliary information associated with the pixels, such as who acquired the data, when it was acquired, conditions of the experiment, etc.
Multi-page TIFF: a TIFF file containing multiple image planes
OME-XML metadata: a block of OME-XML representing image metadata (but not pixels)

The  Open Microscopy Environment consortium has defined a file format called  OME-XML for biological image data including a rich set of microscopy-specific metadata. OME-XML's major strength lies in its extensible, human-readable  XML schemata for image metadata. Unfortunately, OME-XML has a number of drawbacks:

1. Decoding base64-encoded image planes is computationally expensive (though see the  OME-XML FAQ question on performance).
2. There is no facility to split a dataset among multiple files on disk.
3. Far fewer applications support OME-XML than TIFF.

To address these concerns, we have created the OME-TIFF format to maximize the respective strengths of OME-XML and TIFF, taking advantage of the rich metadata defined in OME-XML while retaining the pixels in multi-page TIFF format for greater compatibility.

An OME-TIFF dataset has the following characteristics:

1. Image planes are stored within one multi-page TIFF file, or across multiple TIFF files. Any image organization is feasible.
2. A complete OME-XML metadata block describing the dataset is embedded in each TIFF file's header. Thus, even if some of the TIFF files in a dataset are misplaced, the metadata remains intact.
3. Though our initial focus has been on storing individual datasets as collections of OME-TIFF files, the OME-XML metadata block may contain anything a standard OME-XML file can, including multiple OME images with multiple sets of pixels.
4. The only conceptual difference between OME-XML and OME-TIFF is that instead of encoding pixels as base64 chunks within the XML, OME-TIFF uses the standard TIFF mechanism for storing one or more image planes in each of the constituent files. However, since TIFF is an image format, it probably makes sense to use OME-TIFF only when there is at least one image plane.

For detailed technical information on OME-TIFF, see the OME-TIFF specification. We have a list of software tools for working with OME-XML and OME-TIFF. We also have some example source code in Java for reading and writing OME-TIFF. Lastly, some sample OME-TIFF data is available for download, along with statistics comparing OME-TIFF and OME-XML with various types of compression.