Edward Smith, Scott Fujimoto, and David Meger from Department of Computer Science, McGill University Montreal, Canada are hoping to make some science fiction become science fact.
Imagine that you really like that chair your grandmother used to sit and rock you to sleep in, on the veranda as the sun set over the hills. Maybe you took a photo of the chair when you got that super new camera.
If you were to drop down to McGill University they may just be able to help you get a 3D creation of that chair so you can own one. Publishing their latest paper, they have factored up what was previously possible by 16. Gaining state-of-the-art super resolution results from their new algorithm the team are showing great signs of taking the world one step closer to getting super 3D resolution images from 2D.
With some striking results it’s amazing how accurate the 3D images are that are produced. From Chairs, to planes to vehicles, a simple 2D image creates a fair representative 3D image. What is equally impressive is that the method and all of the computations were run on a single NVIDIA Titan X GPU.