This paper discusses methods of discriminating the shape of an object from a fragment of an actual image and of detecting a target in each image of a video sequence. A set of attributes of the object and a special rule for rewriting the reference – a dynamic measure of closeness – is used to achieve stable tracking.
© 2011 OSA
Original Manuscript: August 23, 2010
Published: June 20, 2011
I. V. Borisova, "Segmentation and tracking of objects on a complex background," J. Opt. Technol. 78, 305-307 (2011)