Jon’s Radio Comments

October 26, 2006

What’s the video threshold for face-reading?

Filed under: Uncategorized — jonsradiocomments @ 2:43 pm

The original item is here.



  1. Microfeature-support is an advantage for recognition-adepts, a disadvantage for others. It is a win for the slender and the handsome, not for the stout and the homely. Cisco’s effort will be doomed by those who will rationally oppose it.

    Comment by John Faughnan — October 29, 2006 @ 2:23 pm | Reply

  2. Jon, I think you are thinking too traditionally by limiting your attention to framerate. The goal is not framerate but recognition which doesn’t necessarily require realtime.

    Imagine a more cinematic approach to video-conferencing [1]. When you are talking, everyone sees a close-up of you. As you are talking, I twitch my face a bit because I strongly disagreed with something you just said. My computer can detect that then signal to everyone’s computers. Each computer then inserts a cutaway to a picture or short clip showing my twitched face as it is done in countless conversation scenes in movies. Usually, someone else will react with another micro-expression which means everyone will see a few cutaways in sequence before the view switches back to your face. Note that all this doesn’t require high framerate.


    Comment by Don Park — October 29, 2006 @ 10:19 pm | Reply

  3. Some more details which I think are relevant:

    While cinematic video-conferencing allows every participant to share a conferencing experience so that there is less chance of some not noticing my twitched face, advanced versions should allow personalized experiences to reflect interest. In a negotiation meeting, micro-expressions of opponents will be more important that that of colleagues so the user should be able to prioritize participants by order of importance which comes in handy at the implementation level.

    Video-conferencing software should also support multiple secure channel of communication to allow participants on the same side to communicate while the meeting is going on.

    Comment by Don Park — October 29, 2006 @ 10:42 pm | Reply

  4. Hi Don,

    It’s a good point. The language of microexpressions is a kind of event-driven protocol. If you can detect an event and code it up, all that really should need to be sent is that code. Potentially a /very/ low-bandwidth solution. Gladwell’s language is evocative:

    “you’ll not only flex your zygomatic but also tighten the orbicularis oculi, pars orbitalis”

    Is it easier to detect these events reliably, though, or just brute-force the problem with a minimum framerate threshold? Great research question…

    Comment by Jon Udell — October 30, 2006 @ 2:59 am | Reply

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

Blog at

%d bloggers like this: