An Italian research team is setting off an autonomous van from Milan heading towards Shanghai, China. Indeed, a very ambitious project. But I am really interested in seeing how it will handle crowded expressways and aggressive drivers in Shanghai. And at the same time, beware Shanghainese!

Source: CNET

Face recognition feature has been appearing in popular album management software/services, this article did a detailed comparison among Google Picasa, Apple iPhoto and Windows Live Photo Gallery.

It seems that Google's Picasa, supposedly backed up by Neven Vision technologies, has the best recognition accuracy, see the screenshot below:

It seems that facial recognition technologies are becoming more and more appealing for the advertising industry. We have known products from NEC and Quividi, now more companies are joining the game, e.g. TruMedia Technologies and Studio IMC. The gender and age aware signage systems they sell are able to target specific ads to certain audiences - so called “proactive merchandising”.

Such technologies are not yet as mature as people may expect. For example, Quividi's system is said to be accurate 80-90% of the time for gender; it can only categorize age in broad ranges — teens, younger to middle-aged folks and seniors; and they admit that determining ethnicity is even more challenging.

Anyway, it is a rising market and we must keep an eye on it.

Read more from Newsweek.

This is somewhat off-topic.

Today I got an ad message from Newegg titled “Experience Enhanced Computer Vision”. Of course it immediately aroused my curiosity.

Well, it turns out to be not the Computer Vision as we researchers expect. What they promote are the glasses from Gunna Optiks that improve human vision when watching computer monitors. What a good title!

HeadMouse 2, developed by Robotics Group at University of Lleida, Spain, is (yet another) face tracking based HCI software. As a direct application of face tracking, similar systems have been existing for a while, e.g. Nouse. My colleagues also worked out similar systems, although not made public.

At CES 2009 Toshiba demonstrated its 3D interface based on gesture recognition technology. Looks cool, but doesn't impress me much (we have seen too many Minority Report style interfaces, and honestly this specific implementation doesn't seem very appealing).

Actually they released the technology earlier in Qosmio G55 (right video), the world's 1st laptop featuring Cell processor technology.

The gesture interface doesn't seems to be fast and natural enough, and I believe there's still a long road ahead before people are ready to use gesture to control their computers and TVs.

We know that face recognition technology is behind Google Picasa's Name Tag feature, letting you group pictures in your web album by people in them.

Now there is a Chinese start-up Tuyuan.com (图缘) offering a similar service, can they grow into a match to Picasa, at least in China?

NEC showed off a robot ticketer at the recent iEXPO 2008 trade show in Tokyo. The life-sized android is equipped with facial recognition technology, and can recommend specific attractions based on the customer’s apparent age and gender.

[Source]

A website called Mgestyk is about to sell a gesture based computer control kit. The sensor is “an affordable 3D camera”, and from the video they posted, their gesture recognition software is pretty robust.

In a recent press conference in Beijing, Sony demonstrated the “Smile Shutter” of its Handycam HDRCX12E, which automatically captures moments of smiles. Face detection and smile recognition are clearly the techniques under the hood. The video below is in Chinese, also be patient if it loads slowly.

Smile shutter is not completely new, it was already integrated in Sony cameras about one year ago.