The Stanford NLP Group released a Named Entity Recognition software, based on Conditional Random Fields and implemented in Java.

It is pretty fast and also acheives quite good performance with the included models.

For integration into IBM's UIMA text analysis frameword, I have written an Analysis Engine component that wraps the Stanford NE Regonizer.
You can download it here: stanford-ner-uima.zip
Just like the Recognizer itself it is licensed under the GPL.

Please let me know if it is useful for you.

The Debian packages of the Java development environment Eclipse normally use GCJ as their Java runtime. However, GCJ is not compatible to the Eclipse plugins the IBM's UIMA framework.
But they work with Sun's Java runtime (isn't this "write once run anywhere" stuff just great?)

So this is how you setup the Debian Eclipse packages to use Sun's Java runtime:
Install the sun-java5-jre/jdk Packages from Debian's non-free archive. This is easy, but Eclipse still depends on GCJ, und uses the GCJ runtime. You could set JAVA_HOME to the Sun JRE (good to try, if your plugins are happy now, look at /usr/share/doc/sun-java5-bin/JAVA_HOME) but we want a more permanent solution:
The Debian Eclipse looks in /etc/eclipse/java_home for a list of possible Java installations. Add /usr/lib/jvm/java-1.5.0-sun to the top of the list.
(and, though probably not strictly needed, you might want to set your system JRE to the Sun one, too: see /usr/share/doc/sun-java5-bin/README.alternatives)