This paper introduces an affine invariant of trapezia, and the explicit constraint equation between the intrinsic matrix of a camera and the similarity invariants of a trapezium are established using the affine invariant. By this constraint, the inne

we present a new method of synthesizing novel views from the vir- tual cameras in multiview camera confi gurations for 3DTV system. We introduce a semi N-view & N-depth framework in order to estimate disparity maps effi ciently and cor