Current images on the web are limited by the amount of work required to edit, label, and program how and where they are displayed. Searching images on the web is limited since it is also bound to manual tagging. Most images on the web only have between three and six tags, all of which are limited by a person’s knowledge and vocabulary.
Look at the butterfly image. Do you know which species of butterfly it is? What about the flower? Not only could Eyegorithm technology segment these image elements apart from the original in a matter of seconds, but Eyegorithm technology can populate the image with meaningful and detailed tags, which will improve its overall use.
See the different information layers of the butterfly image here
Malcom Gladwell, in his book “The Tipping Point”, describes how an idea reaches critical mass and becomes generally known. Such a tipping point was reached by the phrase “fear the beard”. Brian Wilson, a pitcher for the San Francisco Giants, sported a thick dark beard during games that culminated in the Giants winning the 2010 World Series. Local restaurants created food art and artists designed T-shirts. What makes searching this tipping point interesting is that the concept is visual rather than textual.
How can we search without text?
With picture search! Because Eyegorithm’s hyperimages enable us to navigate between smaller images within photographs and video, these smaller images can be used as queries for a visual search. Eyegorithm’s vision processing system integrates with artificial intelligence templates so that it performs standard object recognition, labeling and searching on each layer within a hyper-image. This powerful combination of template technology and Eyegorithm’s hyperimage technology is the power behind this visual revolution. Below is a hyperimage of a photo labeled “Brian Wilson’s Beard”. Notice the visual similarity between images and then imagine accessing any or all of the images simply by selecting a specific image within the menu. Or view other examples around baseball and entertainment by clicking here.
Fans wishing to find restaraunts celebrating “Fear the beard” fever, can locate the sushi restaruant with the bento box show below.
Another example: say fans wish to buy “fear the beard” T-shirts for the game. They use the face close-up (layer 2) to initiate a visual search and type in the word “T-shirt”. A local vender, who used hyperimage images on their website, content can be matched against the second layer, enabling the fan to find the T-shirt seller.
Fans can find and share images of themselves wearing “fear the beard” masks. Each layer can be shared separately and used in memory scrapbooks and collages.
As another example, a website catering to baseball fans, may allow viewers to upload hyperimaged video. This in turn helps the website tap into the social networking activities of its users. The embedded linking features allows the website to maintain targeted ads in the hyperimage display as users copy and send it to each other. In the example below a fan takes a freeze frame of the hyperimaged video, adds his or her own opinion via text layers, and proves that Tim Lincecum recently adjusted his throwing motion.
In an age where content can be copied and shared on the Internet, content producers struggle for new sources of revenue. Hyperimaged movies provide revenue opportunities in two ways. First, hyperimaged movies are fun to navigate, giving users an incentive to watch the content as displayed by the content owner, as opposed to watching a pirated copy. Second, embedded advertising within hyperimages, travel with the hyperimages as they are copied and shared. Enabling content producers to obtain advertising revenue even when viewers are watching their content from a pirated source.
Look at a video example for image layering.