is used for this purpose, which
gives the position of object as output. This extracted
position is then used to extract a rectangular image
template (size is dynamic depending upon the dimension
of object) from that region of the image (frame). The
sequence of templates is generated as object changes its
position.