The sole video
for the "3-Sweep" interactive 3D modelling technique reached 1.5 million views on YouTube since it was uploaded earlier this month. By any standard, that's a whole lot of interest.
Despite that kind of attention, there wasn't much information about the people behind the technique. But through some deft Googling and emailing, we tracked down the man behind the impressive technique.
The main developer behind 3-Sweep is Tao Chen
. He's a postdoctoral scholar at Columbia University, and previously received his doctoral and bachelor degrees in Computer Science and Technology, and Physics, from Tsinghua University in China.
There have been a total of five people working on the project, Chen tells me -- he and graduate student Zhe Zhu are developing the 3-Sweep method, while three professors -- Ariel Shamir, Shi-Min Hu and Daniel Cohen-Or -- are providing instructional and financial support to help get it finished.
The plan is to make the 3-Sweep technique available to the public in due course. Whether the tools are released for free or made commercial remains to be seen, although the massive interest in the 3-Sweep video will no doubt play a large part in this decision.
There were questions regarding the accuracy of the method, and whether or not the examples provided in the video are best-case scenarios designed to show the technique at its fullest.
"It is hard to mention the accuracy on percentage," Chen notes. "The success of the method depends on some criteria for the input image, such as clear edge, ideal projection etc. If the image meet the criteria, it probably works, otherwise it may produce artifacts."
Notably, professor Daniel Cohen-Or
at Tel Aviv University is both aiding Chen with his 3-Sweep technique, and producing a side-project of his own along the same lines, which focuses on applying a similar sweeping method for the reconstruction of 3D shapes from 3D point cloud data.