Face Tracking

The application context for face tracking is a mobile robot that interacts socially with users.

Robust tracking methods have several potential uses in the context of a mobile robot. I'm initially interested in interacting with users, keeping a user's face well-centered in the video frame, and in doing offline learning that depends on well-localized face tracking.

 

Summary

The specific tracking goals for this project are

  1. Robustly follow a user's face in real time, even when the robot (and therefore, the camera) is moving
  2. Reliably follow a face in a short video clip to extract face samples for learning
I'm currently using a method I've dubbed QTrack because it's fast. In my tests, it was about 20% faster than Camshift, Intel's fast face tracker. QTrack uses an adaptive updating procedure, which makes it substantially more robust than Camshift. I don't have a write-up for the method yet, but I have a downloadable win32 version here.

Previsiously, I was using two tracking methods: 1) Bayesian Mean Shift for real-time tracking, and 2) Birchfield's head tracking method for offline processing to learn a face model. These methods are described in my old project notes, which are here. Qtrack is both faster and more robust than either of these, so it supersedes both of them.

 
Home | Research | Computer Vision