A team of Google researchers have developed an advanced image classification and detection algorithm called GoogLeNet, that can recognize a wide range of objects.
Google says in a blog post that ‘GoogLeNet’ algorithm can quickly recognize objects in images and can also locate and label multiple object sizes in one image. The software can even determine an object within or on top of an object. The internet giant says that the technology advances are "directly transferable to Google products such as photo search, image search, YouTube, self-driving cars, and any place where it is useful to understand what is in an image as well as where things are."
The algorithm was created by Google interns Wei Liu and Scott Reed, along with Google researchers, Yangqing Jia, Scott Reed, Pierre Sermanet, Dumitru Erhan, Drago Anguelov, Andrew Rabinovich, and software engineer Christian Szegedy. The software was recently placed first in the ImageNet large-scale visual recognition challenge (ILSVRC). Google has made the software open to developers to further increase its accuracy.
Recently Microsoft's research team also announced a breakthrough in Object recognition technology under Project Adam. Under the research Microsoft voice assistant Cortana will receive the ability to identify a wide range of objects. The technology was announced at the company's Faculty Summit by Microsoft Research's executive vice president of technology and research, Harry Shum. At the event Cortana was successfully able to recognize several dog breeds from just a photograph. Microsoft says that the technology could be used in building new tools for visually challenged people.
Source: Google Research
No comments:
Post a Comment