FB pixel

Android camera APIs must preserve facial appearance for biometrics

The latest Android Compatibility Definitions are here
Android camera APIs must preserve facial appearance for biometrics
 

Android camera software APIs will not be able to alter facial appearance, Google revealed in its latest Compatibility Definition Document, presumably to ensure that biometric facial recognition will work for a range of use cases, reports Android Central.

The paper, intended to set the rules for companies that want to make Android hardware, outlines a series of new requirements that must be met in order for devices to be compatible with Android 11.

The lengthy blog post examines a variety of points related to various aspects of hardware compatibility, but the most interesting one for our biometrics’ audience is the behavior for camera-related APIs.

In fact, the section C-0-12 of the paper openly states that all camera-related APIs “MUST ensure that the facial appearance is NOT altered, including but not limited to altering facial geometry, facial skin tone, or facial skin smoothening for any android.hardware.camera2 or android.hardware.Camera API.”

While the document does not expand further on this section, it is safe to assume it refers to face recognition software in relation to features such as face-unlock, where an initial image is taken by the camera and then used as a match for biometric login.

The criteria set in Google’s latest definitions are facial geometry, facial skin tone, or facial skin smoothening, all elements that help define a specific image.

However, AI cameras often use beauty filters to artificially “beautify” photos. This may cause mismatches between an original photo and a beautified version of it.

Since the old camera APIs are now officially deprecated, smartphone manufacturers will have to develop new ones that comply with Google’s latest Android Compatibility Definitions.

In order to do so, new camera APIs will have to necessarily apply these filters only after taking a photo, as opposed to doing so while capturing an image.

Article Topics

 |   |   |   |   |   | 

Latest Biometrics News

 

India scales farmer ID system for payments with KPMG support

The India office of influential accounting firm KPMG has explained how it supported the advancement of the country’s Digital Agriculture…

 

Digital ID systems fail migrants due to policy gaps, Caribou finds

A new report by research organization Caribou has warned that digital ID systems around the world have continued to deepen…

 

Hopae launches eIDAS 2.0, AMLR onboarding readiness tool

Hopae has launched a free self-assessment tool to help financial institutions offering customer onboarding and identity verification to evaluate their…

 

Certainty vs flexibility – does the UK need a Biometric Surveillance Act?

By Professor Fraser Sampson, former UK Biometrics & Surveillance Camera Commissioner Last week London became a city of two tales. Two…

 

TestMu AI releases testing tool for agent-produced code

TestMu AI (formerly LambdaTest) has launched Kane CLI, “a new browser automation tool that runs directly from the terminal,” and…

 

Travel biometrics making new connections

Airport biometrics projects and companies are breaking new ground and intersecting with other industry trends, from digital wallets to biometric…

Comments

Leave a Reply

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

Biometric Market Analysis and Buyer's Guides

Most Viewed This Week

Featured Company

Biometrics Insight, Opinion

Digital ID In-Depth

Biometrics White Papers

Biometrics Events