The company has filed a patent for the automatic detection of large-scale forms in the images. Soon the engine will be able to search for content to identify them in the media source, without the user having to enter it in writing in the title. An idea that has already aroused the concern of associations for the protection of privacy
A MOUNTAIN VIEW seen along. And above all, they want the web to learn by just watching what is loaded in the millions of videos each day reach the Net in order to identify what’s in the pictures, although the author has not specified in the title, in the description or tags. For the user, what changes? Nothing from the point of view of interaction, within just a short time, Google will be able to search for objects in the videos and photos.
The company has filed a patent for the algorithm ‘Automatic large scale video object recognition’ (automatic recognition on a large scale video objects). The first content to pass under the ‘lens’ of the new program should be videos on Youtube, platform owned by ‘Big G’. The system should then add ‘tags’, that is, for example, labels that indicate the presence of a cat, or a particular monument, which would enable to research not only the title, but also on the body of the video.
All ‘start should be 50 thousand keywords loaded. But the algorithm has aroused the concern of associations for the protection of privacy. A worrying them, it is mainly a possible expansion: “Once users lose control of what happens with their data,” he told the Daily Mail Nick Pickles, director of the NGO Big Brother Watch . Pickles continues: “For example, if a video was added to the information
the place where it was shot, perhaps through the recognition of a famous monument, Google would be able to associate the identity of the person who uploaded a lot more information, such as where he was or if the animals. This kind of power should not be allowed without your consent. “