Researchers Find Use for Smartphones in 3D Model Creation

Researchers at Carnegie Mellon University in Pittsburgh have discovered they can use a simple smartphone as a measurement device in the creation of 3D models. By moving a smartphone from side to side or back and forth while either using video or simply switching the phone's screen orientation from portrait to landscape, the user can get a sense of an object's real size.

While video from smartphones and other devices have been used in creating 3D models for some time, this is the first time they have also been used to determine the scale of models — by taking advantage of relatively inexpensive inertial measurement units (IMU).

"We've been able to get accuracies with cheap sensors that we hadn't imagined," said Simon Lucey, an assistant research professor at the Carnegie Mellon University Robotics Institute, where the research on using smartphones in 3D model creation is being conducted.

For instance, Lucey pointed out, with a facetracker program, they have been able to measure the distance between a person's pupils within half a millimeter, making it possible to go virtual shopping for eyeglass frames.

More than 30 years ago, researchers at Carnegie Mellon pioneered the use of 2D images or video to build increasingly accurate 3D models, but taking scale into account has always been a challenge.

But now, as smartphones incorporate higher frame-rate video cameras, the accuracy of the technique will become even better.

"With a high frame-rate camera," Lucey said, "we can excite the IMU by moving the phone faster, without corrupting the image."

Eventually, he said he believes, similar low energy-consuming techniques will be used with robots and self-driving cars rather than the power-hungry technologies like radar and laser rangefinders now being experimented with.

About the Author

Michael Hart is a Los Angeles-based freelance writer and the former executive editor of THE Journal.

Featured

  • abstract malware attack

    Report: AI Ransomware Can Now Run Much of an Attack Without Human Help

    Ransomware attacks are becoming more automated, and AI could take that much further, according to global cybersecurity company NCC Group's July 2026 threat intelligence report.

  • AI artificial intelligence interface over laptop keyboard

    New Microsoft AI Code of Conduct Emphasizes Human Control

    Microsoft has published a draft code of conduct that details how its homegrown AI models should behave, with human control taking priority over model capability, autonomy, and even task completion.

  • Image of digital screen with data and math formulas on black background

    AI Giants Report Advances in Mathematical Research

    OpenAI has published what it says is an AI-generated solution to the Navier-Stokes existence and smoothness problem, one of mathematics' seven Millennium Prize Problems, while Anthropic has reported separate advances in research mathematics and formal proof generation.

  • man viewing holographic display of data and code

    Why Cross-Skilling Is the Smartest Investment a Campus IT Team Can Make

    Higher education IT teams are being asked to keep pace with AI, cloud, and cybersecurity on flat budgets, and hiring a specialist for every new need can no longer keep up. Building adjacent skills across the people already on staff creates more agile, resilient teams that move faster and retain talent without adding headcount. Pluralsight makes the strategy work through skill assessments, tailored learning paths, hands-on labs and sandboxes, and microlearning that fits into the flow of work.