By David Ponce
Quite some time ago, researchers tried to get machines to render 2D images into 3D. They failed. That was because ‘puters at the time flubbed it. But hey, it’s like 30 years later, and we got some serious chips at our disposal. So, the guys at Carnegie Mellon got to work and have made some remarkable progress. If you come inside, you can see footage of what their program is able to do. Give it a 2D image, and it’ll automatically find visual cues of its composition and extract three dimensional information.
Yeah, I bet there are those of you that will poke a thousand holes into this. Not true 3D, etc. Still. Check out the video, and come back and tell me this isn’t some impressive progress, especially keeping in mind these are the first successful steps in this field.