I have problems in projecting the 3D world points into 2D image points using the camera matrix. I will greatly appreciate your help on this. I simply multiply the camera matrix (C) into 3D homogeneous coordinates (y=[y1 y2 y3 1]') associated to the occupied voxel of the visual hull to obtain x=[x1 x2 x3]'. After normalizing x, i get xn=[x1/x3 x2/x3]=[xn1 xn2]. Then given m=max( max(xn), min(xn) ), I linearly scale the obtained coordinates into the domain 1:64 to obtain the final pixel coordinates of xs=[63(xn(1)+m)/2m+1, 63(xn(2)+m)/2m+1]. Setting the obtained rounded coordinates of xs to one, the obtained image is not as expected
Second point is how to participate the orientation factor in this projection?
Thank you very much for your generous time and support.