Google introduces a face distortion prevention algorithm for selfie cameras

A recent study found that 93 million selfies are posted on social media every day, with 10 being posted to Instagram every 10 seconds.That's enough data to push smartphone makers to offer high-quality selfie cameras to stay ahead of the competition.However, most high-quality smartphone selfie cameras currently have a drawback called face distortion.

Now Google researchers have teamed up with the Massachusetts institute of technology to develop a face-distortion algorithm for self-portrait cameras.The paper, titled "undistorted wide-angle portraits on camera phones," discusses the algorithms they developed and explains how to solve the problem of face distortion in selfie cameras.The following is the research paper on the new algorithm:

Given a portrait as input, we formulate an optimization problem to create a content-aware deformation mesh that locally ADAPTS to the stereo projection of the face area and seamlessly evolves into a perspective projection on the background.Our new power function performs effectively and reliably on a large number of subjects in the photo.The proposed algorithm is fully automated and operates at interactive rates on mobile platforms.We show promising results at various field angles from 70° to 120°.