Comments on: Back to Work No Matter What: 10 Things I’ve Learned While Writing a Technical Book for O’Reilly http://urbanhonking.com/ideasfordozens/2011/07/17/back-to-work-no-matter-what-10-things-ive-learned-about-writing/ Thu, 19 Jun 2014 09:26:37 +0000 hourly 1 By: greg http://urbanhonking.com/ideasfordozens/2011/07/17/back-to-work-no-matter-what-10-things-ive-learned-about-writing/#comment-630 Tue, 19 Jul 2011 17:38:12 +0000 http://urbanhonking.com/ideasfordozens/?p=477#comment-630 To do gesture recognition in a mobile platform you’d need two things: (1) a method for doing depth sensing that was smaller, lower power, cheaper, and worked in sunlight (2) a phone with enough processing power to run the gesture recognition algorithms in real time on your resulting depth image. I think that (2) is inevitable and only a matter of time. The first issue though, presents a much more serious constraint. Right now the Kinect uses AC power, not something made easily portable. It’s IR projector is actually quite powerful and needs this power to achieve any kind of decent range. Finally there’s the issue of working in sunlight. The Kinect’s depth camera works using IR. Sunlight contains significant amounts of IR. Hence the Kinect’s depth camera sees a great deal of interference in daylight, enough to stop the skeleton detection algorithms from operating successfully in many cases.

All of these add up to I think we’re along way away from this technology being easily made mobile. Further I’m not sure what the business use case for putting it in a mobile device would be. So far gesture detection has focused on security and gaming, two things that largely take place indoors.

All of that said, though, I think you could do pretty substantial gesture recognition with mobile phones as their cameras and processors get better. The kinds of things we’re capable of doing even now with conventional cameras are extraordinary. For example, check out this face tracking example: Face Tracking. These kinds of algorithms are sensitive to lighting conditions of course, but I still think they could be widely useful, especially as cell phone cameras get good.

]]> By: Dr William Hayward http://urbanhonking.com/ideasfordozens/2011/07/17/back-to-work-no-matter-what-10-things-ive-learned-about-writing/#comment-629 Mon, 18 Jul 2011 02:38:32 +0000 http://urbanhonking.com/ideasfordozens/?p=477#comment-629 I worked for a technical book publisher planning topics. I’ve heard good things from other O’Reilly authors. I’ve been curious of the processing required for gesture recognition of hand motions (reduced skeleton) could fit in a mobile phone processor.

]]>