The Mountain View-based company is now working on an innovative technology that will be able to identify the sex of the persons displayed in a certain picture. Because some of the most searched pictures are the ones concerning the celebrities, Google aims to develop a technology able to identify the sex and label all of them. "Google is using an Adaptive Boosting meta-algorithm to allow it to identify the sex of persons in a photo from a low resolution grayscale picture of the face," Inside Google reported.
According to the same publication, Google already laid the foundation of the search technology because it developed several relevant tests to obtain information about the upcoming function. "Google says it has 80% accuracy with 10 pixel comparisons, 90% accuracy with less than 50 pixels, 93% on a 20×20 pixel image," the blog added.
However, it is not quite clear if the company intends to identify the sex of the persons included in adult photos. I guess all of you agree that Google owns the best search engine on the Internet, but this doesn't seem to stop the Mountain View company from developing even more powerful solutions to conquer the world. Yesterday, the search giant revealed Universal Search, an improved version of its search technology able to search for most formats of information and display them on a single page. Along with several other updates, Google also debuted a new interface for the search engine that was tested several times in the past.
"The best classifiers published to date use Support Vector Machines (SVM); We match their accuracies with as few as 500 comparison operations on a 20 X 20 pixel image. AdaBoost based classifiers achieve over 93% accuracy; these match or surpass the accuracies of the SVM-based classifiers, and yield performance that is 50% faster," the Google researchers sustained according to ZDNet.
According to the same publication, Google already laid the foundation of the search technology because it developed several relevant tests to obtain information about the upcoming function. "Google says it has 80% accuracy with 10 pixel comparisons, 90% accuracy with less than 50 pixels, 93% on a 20×20 pixel image," the blog added.
However, it is not quite clear if the company intends to identify the sex of the persons included in adult photos. I guess all of you agree that Google owns the best search engine on the Internet, but this doesn't seem to stop the Mountain View company from developing even more powerful solutions to conquer the world. Yesterday, the search giant revealed Universal Search, an improved version of its search technology able to search for most formats of information and display them on a single page. Along with several other updates, Google also debuted a new interface for the search engine that was tested several times in the past.
"The best classifiers published to date use Support Vector Machines (SVM); We match their accuracies with as few as 500 comparison operations on a 20 X 20 pixel image. AdaBoost based classifiers achieve over 93% accuracy; these match or surpass the accuracies of the SVM-based classifiers, and yield performance that is 50% faster," the Google researchers sustained according to ZDNet.
