Google recently published a blog post on its latest object recognition technology which has a got a super sense of recognizing instantaneously all the different objects all kept together in one place. This sophisticated detection system is able to identify a lot of objects even if some of them are partly out of sight.
However the concept of automatic identification of object is quite still tricky, but the internet giant is all set to overcome this obstacle with its cutting edge object recognition technology. This was made possible by using the DistBelief infrastructure in which the concept lies in training the neural networks in a distributed manner in such a way that they iterate very rapidly without the need for any external computing power.
This will result in spotting more objects and making more accurate guesses; it can also detect objects in motion. Google makes a point with it by saying that this technological enhancement could help in better image understanding while performing an image search. It could also be useful in the fields of robotics and such similar fields.
This technology could also be used in other Google products like YouTube and self-driving cars and other similar ones where the understanding of depth, image and perception is important and could prove to be a breakthrough in this field. You may check their official blog for more details.