google vision

For us humans it’s very easy to describe the content of a photograph or image, it’s second nature if you will. However for computers this task just a few years ago at least was almost impossible. Until now.

Earlier this week Google announced that its scientists have been busy fevering away on a new artificial intelligence software program which can describe objects and contents of photographs to a never seen before level of accuracy.

Say what you see

The software is able to describe the content seen in photographs in a very human, natural way rather than a coldness associated with artificial intelligence.

Apart from making a search for images easier the software could also have excellent benefits for partially sighted or blind people.

Google developed a dual layered system for its software whereby one layer processes the image to high accuracy and the other formulates natural sounding language. The process is otherwise known as neutral networking which works in a very similar way to that of our own brains computation process. Over time these systems also learn, get more efficient and develop faster translating into better descriptions of images.

Sights firmly fixed on new data

The software is already accurately describing images the first time it lays ‘eyes’ on the them is also helping to further complex object detection, tagging and classification. All this new data means better captions, better alt-text and a more robust, trustworthy search engine.

In one experiment the scientists presented the software with an image depicting two pizzas on top of a stove (pictured below). The software came back with tthe description “Two pizzas sitting on top of a stove top oven”.

 

pizzas
Automatically captioned: “Two pizzas sitting on top of a stove top oven”

Looking to the future

With the positive results demonstrated above, the scientists behind the fledgling technology are extremely keen on ensuring that their efforts come good. Publishing on a Google research blog they wrote;

“A picture may be worth a thousand words, but sometimes it’s the words that are the most useful…”

Research is still on going with more and more complex images being presented to the software with the hope of producing not only better search results and potential cross-over benefits for Google Glass (once/ if released) but also a profound life changing experience for those with visual disabilities.

As always we will keep you up to date with technology news as it breaks and don’t forget to join in on the conversation as well as follow us on Twitter and Facebook for all things Fortune Frenzy!



Leave a Reply

Your email address will not be published. Required fields are marked *



Back to Top
Footerpoint