On Thu, May 21, 2009 at 4:34 PM, Tracy Harms<[email protected]> wrote: > There is a Twitter image encoding challenge posted here: > > http://stackoverflow.com/questions/891643/twitter-image-encoding-challenge > > As I've been very impressed by the applicability of J to image > processing, some may wish to take up this challenge. (I won't be.) > > The challenge has more to do with judicious choice of lossy > compression technique(s) than with other factors.
Given the nature of twitter, I think uploading to tinypic.com will outperform any other approach. Of course, some people will consider such practicality to be "cheating", so a next approximation would be to include a library of parameterized images in the "compressor" and "decompressor" which in turn would find the best template and then fine tune parameters for the best fit (or the reverse). But why go to the trouble of building up a huge parameterized image library for twitter to deal with a solved problem? Note also that the challenge's rating system only grants 1 point for "following all the rules". -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
