You check those three image streams and look for the factors they have in common. Colour the same, they evoke simliar feelings, they have the same animal.

You're simply decoding those three image streams. No need to do another IS for that.

Alex