A new technique for extracting objects from 2D photographs ad turning them into 3D models has won acclaim online.
The "3-Sweep" method involves selecting the boundaries of cylinders and other objects with a series of easy tools, and then allowing the computing to fill in that blanks behind the object.
The video documents how - in the released demos - it is possible to turn a 2D photo into a model in seconds.
The team behind it (Tao Chen, Zhe Zhu, Ariel Shamir, Shi-Min Hu and Daniel Cohe) said:
"Our approach combines the cognitive abilities of humans with the computational accuracy of the machine to solve this problem. Our technique provides the user the means to quickly create editable 3D parts-- human assistance implicitly segments a complex object into its components, and positions them in space."
The idea was presented at Siggraph Asia 2013. Not everyone is sold on the technique, as this discussion from a relevant forum shows but it's an impressive clip either way.