Reinventing Facial Recognition Technology With Realsense-Based Open Source FACEID


The project is as straightforward as it sounds: your lockscreen is unlocked if the camera’s built-in facial recognition module recognises you.

Intel’s RealSense cameras have been an intriguing piece of technology; we’ve seen a lot of clever hacker uses for them, including robots and smart appliances. Unfortunately, Intel did at one time drop the LiDAR and face tracking-specific models from the RealSense lineup. These, it seems, have not been well-liked, and neither have we observed them in breaches. That is, up until now. [Lina] provides us with FaceID for Linux, a practical application for the RealSense facial tracking cameras.

Since Linux is the intended platform, it must integrate with the Pluggable Authentication Modules (PAM) subsystem for authentication, and naturally, a PAM module for RealSense is available to go with it. It is appropriately called pam_sauron.

This module is a good example of how to build your own PAM integrations and a step towards doing that in a different language for once because it was written in Zig, a contemporary C-like language. However, it’s still a fun and self-contained programme for one of the F4XX-series RealSense cameras if you happen to own one. As usual, there are TODOs, like improving the UX and utilising some security features RealSense cameras have.

Since the release of RealSense, we have observed the use of these sensors in robotics and 3D scanning, at least in part because of their compatibility with Linux. The main RealSense lineup was unaffected by Intel’s decision to discontinue the less popular RealSense cameras, so the hacker-favored depth cameras are still accessible for all of our projects. Interested in the technology involved? Here is a breakdown of a laptop-compatible RealSense camera adapter.


Please enter your comment!
Please enter your name here