Condensation tracking through a Hough space
A. French, Steven J. Mills, Tony Pridmore · Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004. · 2004
Recent work has shown tracking groups of lines through the parameter space represented by a Hough accumulator array to be efficient and insensitive to both occlusion and changes in illumination. Previous methods have, however, been less robust in the presence of distractions are a challenge to trackers of all types, but the problem is heightened in the Hough-based approach as short collinear segments give the same response as an equal length connected line. We describe the application of the condensation algorithm to tracking peaks through a Hough accumulator array, producing a tracking technique that is efficient and robust under illumination changes, occlusion and distractions.