Ever wished you could recreate the effect of those neat multi-lens 3D cameras without having to buy the hardware? Lucky for you there’s some cool 3D tech coming out of Stanford called Make3D. The service uses machine learning to go over your shot and recreate depth and perspective in three dimensions.
Once shots have gone through processing, you get a flyover that does a quick back and forth over the rendered scenery, and a 3D environment that you can walk though using the arrow keys on your keyboard. Users can download both of these files to their desktops for later viewing.
I uploaded about half a dozen shots earlier today and only got one to go through the seemingly stringent processing requirements, but the results on a picture of a beach were fast and impressive. The technology’s not quite perfect, but there’s already a huge gallery of user-uploaded images that have been run through the process and come out the other side with an extra dimension that makes them wonderful to explore. You can view pictures on the gallery without having to sign-up, just keep in mind you’ve got to have Adobe’s Shockwave player installed on your browser to get the 3D goodness.
A similar service from Freewebs called fotowoosh has been kicking around since last April, although it doesn’t have a clear front end for consumers to publish their own photos. Also worth mentioning is Microsoft Live Labs’ Photosynth project which creates a 3D environment using a matrix of photos.
Make3D is a project lead by Ashutosh Saxena, who is joined by Min Sun, and Sung Chung along with Stanford faculty member Professor Andrew Ng.
Here’s a video of the tech in action. There are two more after the break.
Original post by Josh Lowensohn and software by Elliott Back
