Intel To Launch TV Service With Facial Recognition By End of the Year 175
MojoKid writes "Despite television being a rather tough nut to crack, Intel is apparently hoping that its upcoming set-top box and subscription service will be its golden ticket to delivering more Intel processors to the living room. The service would be a sort of specialized virtual cable subscription that would combine a bundle of channels with on demand content. So what's Intel's killer feature that distinguishes it from the vast and powerful competition? Granular ratings that result in targeted ads. Intel is promising technology in a set-top box that can distinguish who is watching, potentially allowing Intel to target advertising. The technology could potentially identify if the viewer is an adult or a child, male or female, and so on, through interactive features and face recognition technology."
A telescreen? (Score:5, Informative)
"The instrument (the telescreen, it was called) could be dimmed, but there was no way of shutting it off completely. [...] The telescreen received and transmitted simultaneously. Any sound that Winston made, above the level of a very low whisper, would be picked up by it, moreover, so long as he remained within the field of vision which the metal plaque commanded, he could be seen as well as heard.[...] It was even conceivable that they watched everybody all the time. But at any rate they could plug in your wire whenever they wanted to. You had to live--did live, from habit that became instinct--in the assumption that every sound you made was overheard, and, except in darkness, every movement scrutinized.
Winston kept his back turned to the telescreen. It was safer; though, as he well knew, even a back can be revealing. "
Via http://gutenberg.net.au/ebooks01/0100021.txt [gutenberg.net.au]
Re:The marketing dweeb bastards won't quit (Score:4, Informative)
Cognovision (Score:3, Informative)
I wonder if this has come out of Intel's acquision of Cognovision, and I hope it's better than what they gave developers in 2011. I worked with this product for about a year, attempting to integrate it into our digital signage systems. Not only did it take a dedicated i5 with 2gb of RAM to safely run in the background, it often (as much as 40% of the time) produced the wrong information about a person. Gender was correct about 80% of the time, age was correct in only about 30% of samples, and race was almost never correct (except for very dark skinned folks were alwas called determined to be of African origin).
The result was that we had to write an algorythm to essentially average out samples over a given duration. When we took 500 samples over 5 seconds, dropped the extremes, and averaged the rest we improved the accuracy to nearly 85%, but that is still pretty bad when you consider that automated actions will be taken based on those results (i.e. to play content for YOUR demographic).
I don't really want my TV watching me.
-MoG