Taking contextual ads to a new level, Bellevue-based Eyealike today announced a new ad-serving system that uses image recognition technology to place ads on social networks and media Web sites.
Eyealike says it’s VisualAd system recognizes details in images and videos such as facial features, skin color, hair, gender and age. It also recognizes logos and product images.
The automated system indexes, filters and classifies images and videos rather than relying on keywords for targeting ads, the release said.
“Keywords and profile data tell us a lot about users, but imagine being able to serve up relevant ads to consumers based on facial or physical attributes, or specific objects in images or videos such as Gucci sunglasses, or even logos like Nike,” Eyealike president, Greg Heuss, said in the release.
Its description of the technology:
“Eyealike VisualAd combines smart facial feature recognition and artificial intelligence matching technologies with 3D model-driven analysis. Using patent pending, multi-phase pruning technology — which clips short snippets of visual-based clues in images and videos to quickly, filter, sift and accurately identify images, motion, and facial features. This approach helps websites anonymously use publicly viewable images and videos in a persons’ profile, regardless of whether they are tagged, to ascertain personal information in a private manner that can be used to serve up relevant advertisement.”