Jim, Well, trigrams are handy for modelling word sequence probabilities in NLP, while perceptual hashes are supposed to calculate similar hash values for similar images, and dissimilar hash values for dissimilar images. So the two concepts are only marginally related
Skip Skip Cave Cave Consulting LLC On Mon, Aug 12, 2019 at 10:02 AM 'Jim Russell' via Programming < programm...@jsoftware.com> wrote: > I read (only the english language bits) and had to ask: Does this relate, > for images, to the work you were doing long ago for text involving > trigrams? I’m not sure I ever thanked you for starting me down that path. > (It was productive!) > > > On Aug 12, 2019, at 3:02 AM, Skip Cave <s...@caveconsulting.com> wrote: > > > > Vanessa McHale has an interesting github site using J to implement a > > perceptual hash for images <https://github.com/vmchale/perceptual-hash>. > > https://github.com/vmchale/perceptual-hash > > > > Perceptual Hash Backgroun > > <https://www.phash.org/docs/pubs/thesis_zauner.pdf>d > > https://www.phash.org/docs/pubs/thesis_zauner.pdf > > > > Skip Cave > > Cave Consulting LLC > > > > > >> On Mon, Aug 12, 2019 at 12:05 AM Joey K Tuttle <j...@qued.com> wrote: > >> > >> > >> Vanessa McHale (@vamchale) > >> 8/11/19, 19:07 > >> mean girls: +/ % # > >> median girls: -: @ (+/) @ ((<. , >.) @ (-:@<:@#) { /:~) > >> mode girls: ~. #~ (= >./) @ (#/.~) > >> > >> Download the Twitter app > >> > >> > >> Sent from my iPhone > >> ---------------------------------------------------------------------- > >> For information about J forums see http://www.jsoftware.com/forums.htm > >> > > ---------------------------------------------------------------------- > > For information about J forums see http://www.jsoftware.com/forums.htm > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
