There's been a lot of good advice given here but I'm going to add my 2 cents anyway : )

I'm doing something similar with parsing HTML on my own since I'm looking at some very different HTML pages so I've had to build my own rules, but anyway...

For your problem I'm going to assume that you're looking for one specific part of your stream.

$foundfile =~ m!(<img src="/ your path here /[^>]*>)!i;
$foundfile = $1;
[download]

This (untested) bit of code should match all that you're looking for and being that you're in a stream, unless you dump that stream in to an array and process it line by line. However if you're looking for every occurence of a graphics format file, change !i to !ig and that should do the trick.

I think that you could probably eliminate the entire path in the stream as you're just looking for <img src=" "> as well.

Hope that helps!

Some people fall from grace. I prefer a running start...