FB pixel

Apple patent application describes Face ID biometrics with different angles, partial faces, poor lighting


A patent application from Apple for technology that would enable Face ID biometric matching with partial face images and in poor lighting conditions has been published by the U.S. Patent and Trademark Office (USPTO), Patently Apple reports.

Patently Apple notes that all three inventors credited for the ‘Robust Face Detection’ technology are part of the company’s machine learning team.

Face detection with Face ID is most effective in upright orientations in portrait or landscape mode, and rotating images based on other sensor data to enable face detection can be unreliable and take a lot of processor power, according to the application. Similar limitations apply to partial faces, which are typically rejected, and both too much or too little background light can prevent face detection.

Apple’s invention deals with this challenge by moving the face detection function to the purview of the company’s TrueDepth camera. It then provides a bounding box for the face within the image, which is presumably where the machine learning comes in. That could allow both partial faces and faces with irregular orientation to be detected. While the patent does not describe the impact on the process in challenging lighting conditions, the infrared TrueDepth system should be less sensitive to light than the standard front-facing selfie camera.

The patent application was originally filed in the third quarter of 2018, and has yet to be granted. The company, like many others, regularly files patents to potentially expand on its technology, with an eye-tracking Face ID system an example reported just last week.

Related Posts

Article Topics

 |   |   |   | 

Latest Biometrics News


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

The ID16.9 Podcast

Most Read This Week

Featured Company

Biometrics Research

Biometrics White Papers

Biometrics Events

Explaining Biometrics