I can not understand how it can be captured by a camera moving so many people. When I photographed the traditional way the result would be unusable with a large number of artifacts arising from movement of people. The only way that I can think of is to use multiple cameras. Sorry for my English XD
shooting something like that is alot easier now a days. with the new mask feature in the ptgui beta4 you could do something like this and not even need photoshop anymore.