Bayesian Learning for Efficient Visual Inference

An electronic copy of the dissertation is available here in Adobe PDF format.

This page contains videos that accompany results given in the dissertation above.
All of the videos can be downloaded in MPEG-1 format.
Most of the videos can also be downloaded using the XVID MPEG-4 codec.

Older videos from previous publications can also be found here.

Chapter 1: Introduction

Page 1 Visual tracking demonstration MPEG video
Page 1 Gaze tracking demonstration MPEG video XVID video
Page 1 Manifold interpolation MPEG video XVID video


Chapter 3: Learning to Track: The Displacement Expert

Page 33 Face tracking MPEG video XVID video
Page 48 Tracking against cluttered background MPEG video XVID video
Page 49 Tracking under severe lighting variation MPEG video XVID video
Page 53 Tracking a car MPEG video
Page 53 Tracking a licence plate MPEG video
Page 53 Affine tracking MPEG video


Chapter 4: Semi-Supervised Learning

Page 59 Inferring hand labels MPEG video
Page 66 Inferring gaze direction MPEG video XVID video


Chapter 6: Outlier Detection and Robustness to Contamination

Page 103 Combining the VIM and the VIC for occlusion tolerant tracking MPEG video XVID video


Chapter 7: Practical Applications of the VIM

Page 114 Eyebrow controller MPEG video XVID video
Page 114 Mouth controller MPEG video XVID video
Page 115 Hand mouse MPEG video XVID video
Page 116 Inferring head pose MPEG video XVID video
Page 117 Inferring camera pose MPEG video XVID video
Page 119 One dimensional manifold interpolation MPEG video XVID video
Page 120 "Do as I do" MPEG video XVID video
Page 121 More "Do as I do" MPEG video XVID video