Our paper (cited throughout the text as [Nehab:2011]) reference follows in bibtex format:

@inproceedings{Nehab:2011,
  title = {{GPU}-{E}fficient {R}ecursive {F}iltering and {S}ummed-{A}rea {T}ables},
  author = {{N}ehab, {D}. and {M}aximo, {A}. and {L}ima, {R}. {S}. and {H}oppe, {H}.},
  journal = {{ACM} {T}ransactions on {G}raphics ({P}roceedings of the {ACM} {SIGGRAPH} {A}sia 2011)},
  year = {2011},
  volume = {30},
  number = {6},
  doi = {},
  publisher = {ACM},
  address = {{N}ew {Y}ork, {NY}, {USA}}
}

We use the idea of recursive Gaussian derivative filters from the following paper (cited throughout the text as [vanVliet:1998]):

@inproceedings{vanVliet:1998,
  author = {L. J. {van Vliet} and I. T. Young and P. W. Verbeek},
  title = {Recursive {Gaussian} derivative filters},
  booktitle = {Proceedings of the 14th International Conference on Pattern Recognition},
  year = {1998},
  pages = {509--514 (v. 1)}
}

Read more in the GPU Computation functions module or start at the namespace gpufilter.