Image: MIT News
Image: MIT News

Thanks to a new algorithm created by Google and MIT CSAIL, smartphone photographers may soon be able to capture images free of distractions like fences and reflections.

Many photographers know how to remove small nuisances, like street signs or utility poles, from photos using a program like Photoshop. With the Google/MIT algorithm, the camera does all the work—and, it can even remove large obstacles such as chain-link fences that can’t be easily handled by editing software.

To work its magic, the algorithm requires a short video or series of images to be shot while the camera slightly pans. Reflections and obstructions are detected by comparing the different frames. The result? One clean photo, and one photo of the obstruction that was removed.

In a short video, researchers demonstrate how the algorithm can remove obstructions from a variety of images. In one example, the reflection of indoor objects disappear from a street scene photographed through a building window; in another, a chain-link fence is removed to reveal a clear image of a tiger at a zoo.

See the algorithm in action:

Dig into the data: Read the “A Computational Approach for Obstruction-Free Photography” research paper.

See more before and after photos at Petapixel.

Related Topics

Share your thoughts

Thank you for your comments and for your role in creating a safe and dynamic online environment. MIT Spectrum reserves the right to remove any content that is deemed, in our sole view, commercial, harmful, or otherwise inappropriate.

Your email address will not be published. Required fields are marked *