Try using HTML::TokeParser::Simple or HTML::TokeParser for parsing HTML. It will save you lots of headaches when you run into random issues on any complex regex you cookup for this task.