Late last month, I shared the news of the death of Bryce Bayer, an Eastman Kodak scientist whose filter array breakthrough nearly 40 years ago is now in widespread use, enabling inherently monochrome CCDs and CMOS image sensors to capture full-spectrum color information.
If you haven't yet viewed the video of Ken Lee's (VanGogh Imaging) presentation at September's Embedded Vision Summit, I commend it to your perusal. Lee begins with a quite hilarious story about an on-site audition he did of one of the first implementations of the company's products...running on automated inspection equipment at a hog farm, and used to monitor animal health.
Lytro's light ray-based plenoptic camera technology, which enables post-capture selective focus on any particular depth region of an image, has received periodic mention on this site. Back in October of last year, for example, I covered the cameras' initial unveiling.
Speaking of intelligent image processing algorithms, I encourage you to check out SmartDeblur, a Windows-based utility developed by Vladimir Yuzhikov that's intended for the restoration of defocused and blurred images. The developer writes:
Google didn't evolve the project name when incrementing from Android 4.1 (introduced in late June) to the more recent and latest 4.2 release. However, the newest "Jelly Bean" version makes several notable imaging improvements that will be of interest to embedded vision application developers.
Back in late September, I told you about Movi.Kanti.Revo, a client-side-rendered, HTML5- and web browser-based, and gesture interface-implemented application co-developed by Cirque du Soleil and Google. For the moment, at least, Movi.Kanti.Revo is only supported on conventional computers (along with a limited set of browsers, at that).
The image sensor is a key piece of any embedded vision system design, of course. As such, many of you are probably already familiar with the term "Bayer Pattern," which I discussed in some detail in a technical article published last fall.
Speaking of infrared imaging...researchers Georgia Koukiou and Vassilis Anastassopoulos of the University of Patras in Greece recently published (in the International Journal of Electronic Security and Digital Forensics) the results of a study of 20 volunteers to test sobriety (or not) by means of computer algorithms.
One of the more impressive embedded vision implementations (IMHO) that I've come across, albeit one of the more potentially troubling from a copyright infringement perspective, is Google's book-scanning, de-warping system.