Skip to Content

Software Transforms Photos Into 3-D Models

Photofly uses overlapping photos to create high-detail 3-D copies of anything from bugs to Mount Rushmore.

Ever wished you could take an object in a museum home with you instead of settling for some photos?

Body double: Photofly can build a detailed 3-D model of a person’s head using just 40 photos taken from different viewpoints.

The design software company Autodesk will release free software next week that could turn those snapshots into your own personal replica from a 3-D printer. Called Photofly, the software extracts a detailed 3-D model from a collection of overlapping photos.

“We can automatically generate a 3-D mesh at extreme detail from a set of photos—we’re talking the kind of density captured by a laser scanner,” says Brian Mathews, who leads a group at the company known as Autodesk Labs. Unlike a laser scanner, though, the equipment needed to capture the 3-D rendering doesn’t cost tens or hundreds of thousands of dollars. An overlapping set of around 40 photos is enough to capture a person’s head and shoulders in detailed 3-D, he says.

The software, which will be available for Windows computers only, uploads a user’s photos to a cloud server for processing and then downloads the results. The 3-D rendering can be viewed as a naked wire-frame model of the captured scene or a version with realistic surface color and texture. The colored models can also be shared for viewing in an iPad app, while the underlying wire frame can be exported in standard 3-D design formats for editing.

Models produced from a well-taken set of photos will be spatially accurate to within 1 percent or less, says Mathews, high enough quality to be used for professional design projects. “You could send that model from your photos to a 3-D printing service to physically re-create what you saw, perhaps at a different scale,” says Mathews. In recent years, the cost of 3-D printers and printing services has fallen, with hobbyist machines like the MakerBot and consumer services such as ShapeWays that will print out 3-D models in a variety of ceramics, plastics, and metals.

Autodesk’s is the first consumer software capable of producing models accurate enough for 3-D printing, says Mathews. Similar projects, such as Microsoft Research’s PhotoSynth, and an app based on the same technology that enables a cell phone to convert its photos into 3-D models, only capture 3-D data good enough to add an extra dimension to the content of photos, says Mathews. The same was true of a previous version of Photofly. “Generating accurate geometry from what we see in the photos is far more exciting.”

Photofly runs through several steps to distill an accurate model from a collection of photos. First, it calculates the position from which each photo was taken by triangulating based on the different views of certain distinctive features. Once the camera positions have been determined, the software goes through a second round of more detailed triangulation, using contrasting views to generate a detailed 3-D surface for everything visible.

“This technology and the popularity of cameras and cell phones means there are now a couple billion sensors out there that anyone can use to create 3-D content,” says Yuan-Fang Wang, a computer scientist at the University of California, Santa Barbara, and founder of VisualSize, which is working on technology similar to Autodesk’s.

Wang says the technology has become robust and simple enough for the consumer market, but there are still limitations that may frustrate some people. “An object cannot be too plain, because the software has nothing to compare, or too shiny, and it cannot be moving much,” he says. Because few ordinary users have experienced the technology yet, it is still unclear how people will handle that, or just which applications will prove popular, Wang adds.

Photofly can be used on objects large and small, from bugs to buildings, and can also handle photos from different sources. A video shows a model of Mount Rushmore created from a variety of online images taken by many different people.

After seeing a demo of the technology at the TED conference earlier this year, paleontologist Louise Leakey has been using Photofly in Kenya to capture early human bones at high detail. The models provide her team with a way to collaborate with distant colleagues and to record accurate measurements of specimens, such as the spacing and size of teeth, without actually handling them (see a video of a specimen captured by Leakey).

Autodesk will also explore using Photofly to capture 3-D models of buildings to speed retrofits designed to boost their energy efficiency. “You can take a bunch of photos and very quickly have a model to make the key measurements needed to figure out what needs to be done to make a building greener,” says Mathews.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.