Is there an alternative to Windows Hello for Linux?

If anyone is still searching for a solution, i got my own script working on the same principles as Windows Hello (using the IR camera and face recognition). It's customizable, integrated with the PAM and works basically everywhere: Login, lock screen, sudo, su, etc.

It's available as a PPA on github if you want to give it a spin.

As all of the answers seem to misunderstand the question (as far as I understand :]), and I don't have the rep to comment I will just post an attempt at an answer/help.

In short: The Windows Hello login seem to use an infrared point cloud to get 3d depth of the field/face (increased accuracy in face ID). Since this is basically the same as Kinect, I suggest looking into Kinect projects and libraries.

Since there are no finished libraries to just plugin in and use, I suggest taking an existing Ubuntu face ID module, modify it by adding point cloud library (PCL) to the face identification algorithm.

Ubuntu PAM face ID: I would try something more recent than the seemingly abandoned pam-face-authenticate, such as this pam-facial-auth, https://github.com/devinaconley/pam-facial-auth fork it, and modify the input data to be the point cloud image from IR webcam.

PCL python binding to webcam: http://pointclouds.org/documentation/tutorials/depth_sense_grabber.php

Hope it helps!

Up to Ubuntu 17.10, it seems there is no real alternative to Hello. There are some projects, mostly based on dlib (machine learning), such as this Python script: http://pythonopencv.com/worlds-simplest-facial-recognition-api-for-pythonubuntu-only/

However, none of them seem so far to have been integrated with PAM