No. ;-/
Yes ;-)
Well, sure the question is what you understand as image search.
Handling a image as page and index meta data is possible just the
image parser is requre.
But to have the text around a image for indexing isn't ready to use.
As mentioned I think it is easy to realize, i see two ways.
First write a small map reduce job that parses the html again and
index text around a image link and the alt description into a new
lucene index and than just write a servlet that use this index.
Alternative you can manipulate the html parser to extract the image
also as link but use some more text around the image as anchor text,
that may should be also a very easy solution.
In case I remember right you may need only to extend the html parser
filter extension (some or a similar name) point.
Stefan
