Amanda Knox: What's in a face?

Some choice quotes:

The eyes are not windows to the soul. They are organs for converting light into electro-magnetic impulses.

"We were able to establish guilt," said Edgardo Giobbi, the lead investigator, "by closely observing the suspect's psychological and behavioural reaction during the interrogation."

There are several good insights throughout the article, many of which will probably seem familiar to readers of Less Wrong. The few that stood out to me:

1. Fundamental attribution error and the general tendency to create grossly oversimplified mental models of others (while simultaneously overestimating our model's accuracy).
2. Various observational biases, especially egregious on the part of police and investigators. They were so satisfied with the "evidence" of her facial expressions, which is readily available (under the proverbial streetlight), that they felt this obviated the need for additional investigation. It appears that this led them to seek only evidence that further confirmed Knox's guilt (confirmation bias), rather than considering ways to disprove the hypothesis.
3. Not sure what to call this other than the (not-too-well-established) Dunning-Kruger Effect: the tendency of nearly everyone involved to overestimate their ability to judge someone's guilt based on expression reading techniques in which they may or may not be skilled.