The ImageAnalyzer project is a ready to use Java library to provide you image retrieval functionalities using various implementations. Actual there is a wrapper for the plain SIFT implementation by Stephan Saalfeld implemented. Also there is an implementation for the retrieval backend using a KD-Tree. The KD-Tree library was taken from Simon D. Levy. Please note that I had to modify the source a bit, so you can't use the original source.
The source of the ImageAnalyzer library itself is GPLv2, same as the Java SIFT implementation (but remember that there is a U.S. Patent on the algorithm itself, so be patient where and what you do with that). The author from the KD-Tree just changed the licensing of the source to LGPL.
Now I will show you a quick usage example. Also the JUnit-Test in the source could be of interest for you.
This is the third version (completely rewritten) of a library I wrote already a year ago. It was originally written for another project at university called OLIVANDA. In short this project is a web 2.0 like application for your mobile device for sharing services about real world bookmarks to bridge the analogous gap. The heart of this system is the picture recognition on photos which are delivered by a mobile device. The problem with this thing was, besides some other bugs, that there are sources and libraries brought together which can't be published in that way. So to avoid scraping this as useless I decided to do a rewrite of this functionality and to provide it to anyone who is interested in.
[23. October 2008]
I have finished the first release 0.01 of ImageAnalyzer. Note that at present the sources are also in the jaiwls repository (svn://neotos.de/jaiwls/trunk/imageAnalyzer). And what is more you can find the Java-Doc here.
[04. February 2010]
I translated the C/C++ FAST Corner Detector to Java, now included in the 0.02 release of ImageAnalyzer. It's functionality is not integrated into the rest, but still useful itself. Note that this is probably the fastest plain Java corner detector you will find!