That would match from the way you can see people in the video holding their hands.

"he device houses an additional four front-facing cameras that work with other sensors to facilitate the software’s 3D effects."

I wonder how the battery life of this will be like.

I was thinking that 4 cameras seems a little excessive when you only really need one for head tracking. But maybe it has 4 and uses one at a time - switching between them when/if one is covered.

Well, you need only need 1 camera for lateral tracking, but I believe you need at least 2 cameras for depth tracking. Possibly a third for eye tracking and a fourth for Skype.