I.R.I.S. Working Group

Interoperable Multimedia Retrieval in Distributed Systems

JPSearch - File Format Download


File Format Source Code with GUI
File Format Source Code without GUI

Installation / Utilization

The JPSearch File Format framework comes with two usage scenarios. The first one is a standalone version which can be used via a console or be integrated into a Java application. And the second one is its integration into the Caliph GUI which supports metadata annotation means. In the following, a guideline for both scenarios is given.


As illustrated before, the standalone version comes with a java library which can be used via a console or instantiated in a larger Java application. The two main scenarios are encoding of metadata into images and decoding of the images in order to receive the metadata elements.

Encoding of images

The encoding of images and its assigned metadata can be accomplished by executing the de.global.Main class. The class provides a set of parameters for configuration which are summarized in the following:

-i_encode <file name>
     name of the input file + format ending. Supported formats are PGM (raw), PPM (raw) and PGX
-o <file name>
     name of the output file with ending .jp2
-auth <author´s name>
     name of the author, 'given name' and 'family name' where both information is required
-conf <confident measure>
     only number in the range of [1..15] is allowed
-flag [yes|no]
     flag read only.
-jp (optional)
     JPSearch XML file. (.xml ending required)
-mpeg (optional)
     MPEG-7 XML file. (.xml ending required)
-dc (optional)
     Dublin Core XML file. (.xml ending required)
-trans [jp|mpeg|dc] (optional)
     Metadata format from which the transformation starts.

If no metadata exists, jp2 file won´t be created.

A possible call might be look like as follows:
java -i_encode iris.ppm -o iris.jp2 -auth Joe Doe -conf 7 -flag no -mpeg mpeg7eval.mp7.xml -dc dublincoreTest.xml -trans jp de.portal.Main

Decoding of images

Decoding simply is activated by using the following parameter:

-i_decode <file name>
     name of the input file. it has to be a file with JP2 or JPEG format.

This results in a decoded image in the PPM format and the attached metadata format files of the JPSearch file. Those XML documents (currently JPSearch Core, Dublin Core and MPEG-7) are stored at the local file system. Furthermore, the console shows information regarding the annotation part of the JPSearch file format.

Caliph Extension

By using the Caliph GUI, a user has extended possibilities in annotating image metadata. However, for encoding and decoding functionalities the beforehand mentioned java library is integrated. In the following, the steps for using the GUI are described.

Encoding of images

The first step, after starting the JPSearch Caliph GUI, is the loading of images in PPM format. During the loading phase, the extraction of MPEG-7 information is realized automatically. Afterwards, the user is able to annotate information at the individual metadata formats by using the respective annotation screen. However, beforehand an overall metadata form has to be filled which requests information such as creator, rating, etc. This information is global for all metadata formats.
By switching between the individual metadata formats an automatic transformation among the descriptions is realized. After finalizing the annotation, the information can be saved (using the Create JPSearch command in the menu) into a JPEG 2000 encoded file basing on the JPSearch File Format specification.

Decoding of images

Decoding of images can be realized by loading JP2 images (following the JPSearch File Format specification) with the JPSearch Caliph GUI. The internal metadata is automatically extracted and the existing metadata formats are highlighted through the FileBar-buttons. By using those buttons, the user can switch between the screens of the selected metadata formats.

back to JPSearch