Google patents video tech that knows what it's looking at
The company says its technology can recognize what's in a video clip -- the Eiffel Tower, say, or a simple chair -- regardless of whether users tag them or not.
Google has won an interesting patent related to online video.
Dubbed "Automatic large scale video object recognition," the patent describes a method by which the company's technology will identify objects in a video without the need for the user's input. So, if a person is taking video of a globally famous landmark like the Eiffel Tower or a product like the Porsche 911s, Google's technology might be able to identify it.
In order to determine what different objects are -- both famous and seemingly inconsequential, like a chair -- Google will have an "object name repository." That repository is slated to hold "at least 50,000 object names," the company says.
"Conventional systems rely on direct human input to provide object exemplars explicitly labeled as representing the object, such as a set of images known to include, for example, dogs, based on prior human examination," Google writes in the patent, which was filed back in February 2009. "However, such human input is expensive, time-consuming, and cannot scale up to handle very large data sets comprising hundreds of thousands of objects and millions of images."
It's not clear how Google might implement its technology, but if it does launch, it'll likely be made available on YouTube.
Google has been offering up a host of additions to YouTube lately. Back in March, for example,, allowing users to automatically modify their uploaded clips.