Phase-Based Measurement of Optical Flow and Binocular Disparity

Research Overview

Among my most significant contributions in machine vision are the development and analysis of phase-based techniques for measuring optical flow, image orientation, and binocular disparity. With Allan Jepson (University of Toronto) and Michael Jenkin (York University), my research on local phase information has involved both experimental and theoretical work. In addition to developing algorithms for phase-based matching, we have examined the stability of phase information, showing how unreliable measurements can be identified and rejected. Leif Haglund (PhD, Linkoping University), Keith Langley (University College London) and I have since extended these techniques with the use recursive spatiotemporal filters toward real-time implementations, and to the measurement of multiple orientations in small image regions for estimating surface orientation.

Related Publications

Return to David Fleet's home page.