MSearchGroove

Another sign that multimodal search is finding its niche. Accenture Technology Labs have developed an image recognition technology that works using video taken from a 3G cameraphone. In this scenario Accenture uses an algorithm called Scale-Invariant Feature Transform (SIFT) to identify objects in live video footage. (Put simply, users capture a video clip of an object in their surrounding area, send it to database and then receive a match related to items the system ‘recognized’ within the video.)

A service offered by Magnet Harlequin, for example, uses images users capture with their camerphones to support brand recognition and marketing campaigns. And there are many more lining up to stake their turf in this nascent market.

Indeed, image recognition space is crowding as more companies harness this technology as a means to overcome the usability barrier of mobile phones for search (tiny keypads, tinier screens) and excite the all-important impulse buy. I recently finished an in-depth feature on the topic and will post it soon in the Portfolio section of this website (with another 30 or so articles about mobile search I have yet to list…)

The technology is super-cool, but I’m most excited about the mobile commerce and advertising models visual search could jumpstart. Case in point is GeoVector in Japan. The company recently made headlines when it teamed up with Japan’s Mapion, a mapping services company, to provide a local search service that allows users to look for destinations and information about some 700,000 places of interest across Japan. The new release of the mobile search service, now available on over two million Sony Ericsson, Kyocera and Casio mobile phones using the KDDI network, also enables opt-in advertising, sponsored categories and preferred placement.

Meanwhile, Microsoft is working to make it a reality through Photo2Search, research in progress that uses images to act as the search query, with no text input from the user whatsoever. Users simply take a picture of a given image with their cameraphone; submit it to Microsoft’s search engine, and all the details are delivered directly to the mobile phone. Nokia has also filed a patent that lays the groundwork for a mobile visual search service predictably linked to Nokia cameraphones.

Not to be outdone Google, which acquired Neven Vision, a provider of vision technology in August 2006, also has ambitious plans for image recognition and search. “We believe Neven Vision’s recognition technology offers promising integration possibilities with Google’s services, such as Picasa and Picasa Web Albums, particularly for helping users organize and search their own photos, but we have no specific plans to announce at this time,” a company spokesperson said in a statement.

The interest of major search engine giants signals a new stage in the race to create and monetize a visual index of the world around us. Vendors, services and providers to watch (no pun intended) include DSPV Ltd, Mobot (acquired by Neomedia), Mobile Acuity, SnapNow, LTU Technologies, Polar Rose, Xcavator, and Picapage. And this list goes on (always a good sign that we are talking about a market going places)…If you’re a company in this space, or want to be included on the list, then contact me for a briefing @ peggy AT msearchgroove.com

• Posted in
• Location-based services, Multimodal Search, Mobile Search, Content Discovery

1 comment

share

These icons link to social bookmarking sites where readers can share and discover new web pages.

• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 
• 

permalink