I know you have them, thousands of them. Photos, the 2D type… and even more images you’ve cut out of magazines and library books. If you’re like me, you’re fine with gluing them on odd shaped creatures and giving them out as gifts. If you want something more, you may be interested in 3Defy, a web app offering the ability to push and pull those 2D pixels right out of their flat, ever-loving environment. Why would you want to do that? Well, because you’re a creative type, into online art-forms and you also really, really want to make that Mary Poppins wallpaper really pop off the screen. I leave you to it as Chris Wallace, founder of 3Defy, explains what this tech has over other 2D photogrammetry tools like PhotoSculpt or Autodesk Photofly.

3Defy: Web-based Photogrammetry tool

SS: So this is pulling the pixels off the image? Is there development plans to make it fully 3D or even be able to get usable geometry out of it? That’s one thing I think the photogrammetry tech is missing, although Autodesk has made some leaps with the Photofly project. What are your thoughts on that?

CW: The primary application of our technology will be to generate images and videos that can be viewed on 3D monitors.

Photofly is an awesome technology, but it is only applicable in situations where a multitude of photos can be taken around an object. What distinguishes 3Defy is its ability to create compelling results in situations where only a single 2D source photo is available. We’re hoping to unleash the 3D power of all the years worth of great content that already exists in 2D form.

Most existing tools (including those from Autodesk) appeal only to graphic designers and 3D specialists. Our goal with 3Defy was to create a tool that is easy for anyone to use, making it fun to create 3D models from their photographs. We see it as a new type of online art-form, and we are hopeful that the community will take advantage of the sharing capabilities and the embedding API that we’ve provided.

Regarding the challenges of single-photo 2D to 3D, many researchers have tried to accomplish this with automated algorithms (the Carnegie Mellon team was the most prominent). However, nobody has been able to achieve realistic results, due to the difficulty that Computer Vision and Artificial Intelligence algorithms have in identifying distinct objects and determining their original sizes. The reason that 3Defy enables such realistic results is that it relies on user interaction to identify important objects in the scene and recreate the real-world 3D structure.

Wondering what the new fangled photo web-tech is like? You can check out the Gallery chock full of examples and then have a go at it yourself with their web-based “3D Modeler” that pulls out the pixels as you paint your cursor across the picture.

Author

Josh is founder and editor at SolidSmack.com, founder at Aimsift Inc., and co-founder of EvD Media. He is involved in engineering, design, visualization, the technology making it happen, and the content developed around it. He is a SolidWorks Certified Professional and excels at falling awkwardly.