Research outlines strides in facial recognition through deep learning
Facebook has revealed in a research paper that it has discovered a way to match faces from two different photographs with 97.25 percent accuracy. The research has aided in the development of a software project called DeepFace, which only slightly trails the facial recognition measured average of a human by only 0.28 percent.
Details app that recognises item or media, returns Amazon store page
Rumors of an Amazon-created smartphone have had a boost, thanks to a recently filed patent application. The unearthed filing for "User guided object identification" provides details for how a smartphone can recognize an object and then provide the user with an appropriate Amazon product page, a feature that was mentioned in a rumor from earlier this month.
1000 machine neural network scanned 10M YouTube images
Google's X Lab has built a one-billion-connection "neural network" that can identify cats on YouTube. The project, aimed at simulating object recognition by humans, was able to more than double the accuracy of item identification from a list of 20,000, according to the New York Times, Using 16,000 cores in 1000 connected machines, the system learned to identify objects without human supervision. The technology represents a departure from current vision-learning methods.
Agency agrees with opt-in policy
A German government agency that oversees privacy issues has reportedly approved Google's implementation of facial recognition technology. Although the same agency has warned that it will take action against Facebook for using similar technology, the regulators are said to be less concerned with Google's opt-in approach.
Feature aims to streamline tagging
Google has introduced facial recognition technology for Google+, enabling users to quickly tag photos without manually entering identities for individual pictures. Users can access the new feature by creating a new album with pictures that include people's faces. The system automatically groups photos for each person, with options to tag by name or e-mail address.
iPhone writing recognition
ModMyiFone.com forum members have discovered a new utility for jailbroken iPhones which allows for handwriting recognition using a small grey area. The beta software can recognize both Latin and Chinese characters at the moment, transcribing finger motions into editable text on a yellow notepad-style document. User poetic_folly notes that installing HWPen requires the user to add iphonecake.com/src/all as the installer source.
Apple seeks engineer
Apple has posted the availability of a Handwriting Recognition Engineer position, citing the need for a dedicated employee for its Mac OS X handwriting recognition technology. According to MacRumors.com, the individual would be responsible for enhancing the existing product, as well as extending the technology's influence to "other Applications and the iPhone," suggesting Apple is trying to recapture the feel of its Newton handheld platform.