November 12, 2007 |
Just last week, I covered an innovative image search engine, known as Imagery, that introduces some very welcome features to make the process of searching for images on the web easier.
Even though less clicking is involved and there is no need to visit external sites with Imagery, the basic searching functions the same as all of the other existing image searching options. There is a major flaw in the current approach, according to visual-based search innovator ActiveSymbols, in that today's search models are text-based and therefore ”not naturally designed for image-based search and retrieval because they rely on rich hyperlinks, anchor tags, metadata, path names, font size, and page formatting to deliver results.”
Making the first move towards a completely redesigned image searching model, ActiveSymbols has today announced the launch of its Eyealike Visual Search (VS) platform which goes “beyond the traditional text-based model with a new image-to-image approach that offers unprecedented speed, scale, and relevancy for connecting consumers to the rising tide of digital media.”
According to ActiveSymbols, the Eyealike VS platform is patent-pending and offers the deepest and most automated analysis of digital images and videos, while also being the first enterprise search solution for three key models: facial recognition, image detection, and video copyright surveillance. Where current image search systems depend on text (such as keywords and tags), the new Eyealike platform is able to analyze online objects, delivering unprecedented speed, scale, and relevancy for helping consumers connect to the rising tide of digital media.
?Our core mission is to dramatically advance the way consumers search, find, and access people, images, and other rich media content online, whether they?re at their home or work PC or using their cell phone,? said Greg Heuss, president of ActiveSymbols. ?Text-based search isn?t going away anytime soon, but we no longer live in a world where it will suffice. Visual-based search is the future and our goal is to power Websites that need to connect their consumers to rich media in a more meaningful and powerful way.?
The first module to launch is Eyealike Face Search, which utilizes facial recognition technology. One area where this could potentially be very useful is in the online dating scene, which always has people search though dozens of images, trying to find someone who they are first physically attracted to. Often people search using short descriptions (i.e.: hair color, eye color, skin complexion) that a user may include on their dating profile, while Eyealike's Face Search algorithm will actually be able to analyze a user's photo and evaluate “the complete geometry of the face, including skin tone, eyes, hair color, texture and length as well as hundreds of other relevant data points.”
All in all, ActiveSymbols seems to have created some very promising developments that definitely have mainstream potential. Perhaps I spoke a little too soon when calling Imagery the best advancement yet in image searching.
The Eyealike Image Detection and Video Copyright Surveillance modules are currently in private beta. Those willing to participate can request access by sending an email to firstname.lastname@example.org.