Name: Emotient v5-H264 1280x720 Med (1)
Uploaded: 2014-06-10
Duration: 2 min 44 s
Description: Emotient v5-H264 1280x720 Med (1)

At Emotient, Wwe are the foremost authority. We are the driver, the leader of a new, innovative space called emotion aware computing. We are driving the next generation of personal technology by enabling businesses, people to extract from facial expressions actionable insights that will eventually lead to better products, better service, better user experiences. 10 years ago, the technology was still nascent. The technology for automatic facial expression recognition had not yet gotten to the point were it was real time and robust. Right now, we have those ingredients that we can put together a real time system. Today we can do the processing that's required on one core of a Macbook Air or a similar Window's based device, and even, we, we've run our system on tablets and even smartphones. If you can grab that information in an anonymized fashion, aggregate it, do the analysis, the analytics around it, you have an opportunity for product developers and companies of all shapes and sizes to deliver better value to consumers faster. There's so many things that technology can apply to. It's a very horizontal thought. We decided to focus on retail and healthcare initially. Those are areas where we, we believe we can make, you know, the most immediate impact. So our software is able to detect facial expressions in each frame of video, but then when you do that over time, what you end up is with is a time course, so you get this nice information about the rise and fall of emotion over time. And so for each frame of the content video, you'd, you have information about how the audience responded to that specific point in time. One of the things about our technology that's really powerful is the fact that we can, we can analyze multiple people at a time, multiple people in a frame. The limitation is, is basically, you know, based largely on size of the face in the frame, so, but our spec is roughly 48 by 48 pixels. The fact that we can do multiple people simultaneously is a real strength and obviously provides great leverage to the model. Of course, with the human visual system, if you see a face, you see who it is and the expression that the person has at the same time, but we can design a computer system that actually recognizes the expression without detecting anything about the person's identity. The privacy issue is, of course, very important to a lot of people, including us. This is a new age and it's, it's part of the process of acquiring information, accumulating it, analyzing it and using it in a way unheard of before, and it's different information. It's the information that's available on our face.