You could try:
The webserver will run
getpic.pl, and you can parse bobo.jpg out of the http header "REQUEST_URI". The example above would return a REQUEST_URI
of /getpic.pl/bobo.jpg, for example. Use a regexp to grab the username, look it up in the database, generate an image/jpeg content-type header, and dump the file out.
Update: I think I should explain this a little better.
The web browser will think that getpic.pl is simply a directory, and that
what it is accessing is a file called bobo.jpg. If you return the right headers, the browser will think
that it is retrieving a static jpeg. Depending on how the browser determines whether or not to reload the picture, that may be all that you need.
On the other hand, some browser will request the files headers in order to check modification time, file sizes, and whatnot. You need to be prepared to fake those headers if necessary.