XML::Driver::HTML is a SAX Driver for HTML. 

There is no need for the HTML input to be weel formed, as XML::Driver::HTML is
generating its SAX events by walking a HTML::TreeBuilder object.  The simplest
kind of use, is a filter from HTML to XHTML using XML::Handler::YAWriter as a
SAX Handler.

WWW: http://search.cpan.org/dist/XML-Driver-HTML/