BMVC 2004, Kingston, 7th-9th Sept, 2004
Occlusion Analysis: Learning and Utilising Depth Maps in Object Tracking
D. Greenhill, J. Renno. J. Orwell and G.A. Jones (Kingston University)
Complex scenes such as underground stations and malls are composed of
static occlusion structures such as walls, entrances, columns, turnstiles, barriers,
etc. Unless this occlusion landscape is made explicit such structures
can defeat the process of tracking individuals through the scene. This paper
describes a method of generating the probability density functions (PDFs) for
the depth of the scene at each pixel from a training set of detected blobs i.e.
observations of detected moving people. As the results are necessarily noisy,
a regularization process is employed to recover the most self-consistent scene
depth structure. An occlusion reasoning framework is proposed to enable object
tracking methodologies to make effective use of the recovered depth.