Wednesday, April 29, 2015

Next Generation Facial Recognition Software Knows How You Feel

By: Anthony Germinario

Affectiva, a pioneer in emotional recognition software, seems to be everywhere lately – from discussions in my office about new MR techniques, to a recent article in Wired. I first heard of their Affdex technology at an ASC conference in London, and was thoroughly impressed by the ability to capture emotions while respondents view videos or ads. Facial recognition has been around for a while now (remember when Facebook started guessing who was in your photos?) but decoding emotions on those faces is a whole new frontier.

Affdex was developed with altruistic intentions at the MIT Media Lab; to help autistic people read emotions during daily interactions. A machine that reads emotions, however, inevitably caught the attention of many more interested parties. While I can get excited about using their technology to measure respondent reactions in Market Research studies, I am even more interested to see where else this will be applied in my everyday life as a consumer. Which of my devices will read my emotions, and what will they give me in return?

Affectiva recently offered a 45-day free trial to developers who want to experiment with their API – which got me thinking... what are some apps or devices I would want to read my face/emotions? I’m not a developer (just a dreamer) so here is my short list:

1)      Apple TV / Roku – Could the device please pause my show when I inevitably doze off while catching up with my shows on Sunday evening?

2)      eCommerce sites (Amazon, Gilt, etc.) – While I shop, can you tell which items I react positively to, and tailor my experience like a virtual personal shopper?

3)      Dating sites – maybe Tinder can tell exactly how you feel about a potential match, so you don’t have to keep swiping left/right? Perhaps you would find different matches based on your initial emotional response, which you may not even be aware of.

All dreaming aside, one real concern about any new kind of data capture, especially involving video, is privacy. Consumers are willing to trade a good amount of personal privacy for novelty and convenience, but it certainly is something that must be addressed. Rana el Kaliouby (Chief Science Officer at Affectiva) assures us that Affdex, while it has amassed a database of millions of faces, retains no personally identifiable information. So even if we know my face is in there, she asserts that nobody would be able to pull it out of the system.

I, for one, will take her word for it – and am excited to see this approach applied in consumer technology. What about the rest of you – any other ideas for places you do (or maybe don’t!) want to have your face/emotions read?

About the Author: Anthony Germinario is Director of Technical Product Management at BuzzBack, where he is focused on developing and integrating unique respondent and reporting experiences for online research. He has earned his PMP certification and holds a B.S.B.A from Boston University’s Questrom School of Business. You can keep up with him on Twitter @AGermBB and on LinkedIn, as well as on BuzzBack’s blog.

Related articles

No comments: