If the ability to represent the goatse image were a critical part of the evaluation process for a compression algorithm, then yes. However, we evolved to recognize and respond to subtle features in human faces, not human colons. So, no, goatse would not be an appropriate reference image.