For PDF files the following one liner might help you. I didn't do any decent rearranging of the code:

perl -ne 'undef $/; $\="\n"; 
(($count = $1) =~ s/.*\Count\s*(\d+).*/\1/ && 
$pages < $count ? $pages=$1 : 1) 
while s/(<<[^<]*\/Type\s*\/Pages[^>]*>>)//; 
print $pages' somefile.pdf
[download]

I tested it one some pdf files and it seems to work. It is not a very elegant solutions and might be heavy for large files. It reads out the sections like the following in the PDF file and assumes that the larges count holds the number of pages, which is really an assumption.

23 0 obj
<< /Type /Pages
   /Resources 76 0 R
   /MediaBox [ 0 0 595 842 ]
   /Kids [ 24 0 R
            // and so on
           ]
   /Count 11
>>
endobj
[download]

- wine