2014HS: 31004 Computational Photography

Oh thank you, yes I think I got it. I was confused with the uv/st plane, camera on uv or st plane, but I guess for the basic understanding this does not matter and what also confused me is the location of the origins. If the cameras are on t, one can fix that numbering and for each ray the origin on v changes but not on t..

If someone can use it, I drew a little sketch using different colors which helped me. One shows the two "plane" parameterization and the other the ray space for two points (for rays) to see that the slope reduces for the point that is further away. Key was to fix the origin of t and for each ray determine the new origin for v (also shown using different colors). Maybe someone can use it..