Abstract: Systems and methods for generating efficient transformed input image address sets for producing a multi-pane output image from an input image are disclosed. The input address sets may be generated by applying a first transformation corresponding to one pane of the output image to output pixel addresses to create first transformed input addresses, applying a second transformation corresponding to another pane to the output image pixel addresses to create second transformed input addresses, and storing, for one output pixel address, a first transformed image pixel address and a second transformed input pixel address and, for another output pixel address, including a first transformed input pixel address, but no second transformed address.