The cognitive levels of vision problems can be a challenge. Computer vision researchers tend to think that the problem is easier than it is, often underestimating the complexity of higher level processing.