Signal Processing and Communications Laboratory

Department of Engineering

Pose estimationMichael Burke

Background - Research - Publications

Position: PhD Student

E-mail: mgb45 [at]

Office Location: BN3-02

Thesis Title: Pantomimic gesture recognition for human-robot interaction

Supervisor: Dr Joan Lasenby


I am currently on a leave of absence from the Council for Scientific and Industrial Research, South Africa, where I have worked as a field robotics researcher since 2009. I completed a part time Masters degree in Electronic Engineering at Stellenbosch University while working at the CSIR, with my thesis on visual servo control for a human-following robot. Prior to this, I obtained a bachelors degree in Electronic Engineering from the University of Pretoria.

Research Interests

Statistical methods applied to computer vision and pattern recognition, with applications in mobile robotics and motion capture. The videos below show selected results on pantomimic gesture recognition using robot behaviour recordings as training data and single camera human pose estimation. Our work on the latter recently won the best paper award at the VIII Conference on Articulated Motion and Deformable Objects (AMDO 2014).


Google scholar