Contributing images with metadata
The documentation for both the current version of the Midas XMP editor and the metadata element set defined for MIDAS can be found here. This documentation contains all information provided in this section in a pdf document.
Background
In its basic form, an image file is identified by nothing else but its file name. Typically, the file name has an ending (suffix) such as ".bmp", ".jpg", ".tiff", etc., which tells us (and the computer) which type of software is required to view the image.
If the image file has been produced with a digital camera, then there will be some information stored in the file along with the image data. This so-called metadata can be read and displayed by most newer image processing programs. Cameras use a metadata set named EXIF to record the date and time, exposure settings, and the maker and model of the camera.
Most digital image processing software also allows the user to enter a title or caption, keywords, and other content-related information, to be stored inside the image file.
There are various ways of storing content-related metadata within an image file. What has been entered with one program cannot always be retrieved by another. Therefore, interoperability between image processing programs is still an issue. Moreover, existing metadata specifications have been developed by specific groups of users for a limited area of application. The IPTC (International Press Telecommunications Council) metadata specification is usually adequate for news photography, but will not be suitable for medical images such as X-rays.
Since there is no single metadata standard for all application areas, a "container" format has been devised that allows for the aggregation of metadata elements from different standards and specifications. Named XMP (for Extensible Metadata Platform), this specification is currently being implemented by an increasing number of image processing software manufacturers.
XMP in the MIDAS environment
Still images can be an important information element for users of film catalogues. Therefore, MIDAS encourages archives to supply still images relating to catalogued films wherever possible. Arranging a large number of image files for display via the catalogue server can be a time-consuming and error-prone task. For this reason, we ask all suppliers of still images to mark the corresponding image files with a small set of metadata elements. The most important of these elements will be the MIDAS film work identifier, which will allow the catalogue server to ensure that the image is displayed in combination with the right catalogue record.
Since the required metadata elements are highly specific to MIDAS, we have decided not to re-use existing specifications, bu rather to define our own and wrap these up in an XMP metadata container. This is fully conformant with the XMP specification.
Documentation of the MIDAS metdadata elements can be found by following the link at the end of this article.
Preparing still images for upload into the film catalogue
Currently, there are the following methods for preparing image files for upload to the catalogue server:
1. Use of a stand-alone program for Microsoft Windows that allows MIDAS-specific metadata to be created, edited, and saved in several types of image file.
2. Use of other tools such as text editors for writing the XMP-formatted metadata into text files, and of tools for writing blocks of XMP data into image files.
A first version of the Midas XMP specification (issued in 2008) also recommended the use of the Adobe Photoshop File Info panels as a data entry tool. This recommendation has been withdrawn for several reasons. One of these is that the customisation procedure for the File Info panels differs widely between versions of Photoshop (with no support at all for versions before CS2), thus MIDAS: Contributing Images with Metadata 3 requiring an inordinate amount of application support. Another, more critical, reason was the incomplete implementation of RDF data structures at the interface between Photoshop and the XMP software component. While Adobe's reference implementation (published as the XMP toolkit) correctly handles all RDF constructs defined in the XMP specification, Photoshop software products restrict the object part of any RDF statement to a simple, monolingual literal.