Using your face to unlock your smartphone, or seeing Google or Facebook’s algorithms identify your friends and family proves that facial recognition tech can be useful. But where there’s a will, there’s a way to… repurpose technology for the adult entertainment industry. On webcam show hub Megacams, you can upload a picture of yourself and it uses machine-learning facial recognition magic to find your sex-cam lookalike. I’m not sure why you’re looking for someone that’s pleasing themselves on camera when they look just like you. Regardless, the technology is here — although the site is (obviously) not at all safe for work-time browsing. I warned you.
Out of morbid curiosity. I sent my photo into the site’s webform to see if I had any exhibitionist clones out there. The summary: nope. No-one looked anything like me, and I’m not being subjective — or at least don’t think I am.
This is me:
Wholesome, fully clothed, not-smiling Brit that I am. Disregarding the lookalikes that had their junk out, here are my apparent doppelbangers:
The percentage match here was 46, 42 and 40 percent respectively, with other 40-percenters looking even less like me. And let me repeat: these people do not look like me either. The biggest similarity here is that I’m showing a bit of arm (risque!) as seen in Doppelgangers #2 (“fat gamer Mat,” according to a colleague) and #3.
The site isn’t revealing exactly which face algorithm tech it’s using, although TechCrunch points to Microsoft’s Face API recognition software as a possible culprit. There’s no shortage of options: it could even be using open source tools like openCV, used in everything from cars, to smile-spotting apps, to virtual pottery wheels.
Natasha Lomas at TechCrunch had a similar experience with her matched webcam artists, noting most of her webcam “twins” that scored under 40 percent looked nothing like her. Results are more disheartening than Microsoft’s age-guessing bot getting it wrong. I think I’ll tape up my webcam. Forever.
Via: TechCrunch
Source: Megacams