APPENDIX
(Equations)
Equation 1:
	p (x, b ) = Ii (x, Ci (x, Bi (x, b )))   (1)
	where x is a parameter vector calculated for optimum alignment of the projected model with the image, B is the Body frame of reference transformation, C is the Camera frame of reference transformation, I is the Image frame of reference transformation, b is a body-part surface point, p is a pixel in 2D frame of video (Rehg and Kanade, 1995).
Equation 2:
	edge error Ee using edge information:
Equation 3:
	region error Er using region information (Equation 3):
where it represents the image at time t, mt the model gradients at time t, ne is the number of edge values summed, ve is the edge variance, nr is the number of region values summed, vr is the region variance, pj is the image pixel coordinate of the jth surface point on a clone-body-part.