From the smartphone's camera it recognizes objects and offers related results: here is Blippar, a database larger than the Internet itself
If he kept his promises, this system could revolutionize the concept of search engine. The London startup Blippar has created a "visual browser" that uses the machine learnig to recognize objects from the real world.
By accumulating recognition of similar objects, Blippar's proprietary system analyzes the characteristics of an object but also of plants, animals, logos for suggest what it could be.
Just point the camera at theiOS or Android app and give time (little, to be honest) to process the information. Related words will begin to appear first, tag cloud model, and after a while, when the application is secure, put the result of its analysis at the center. If there were more than one, they will position themselves in a window at the bottom, in a row. By clicking on it, you access the results tab. Above the meaning, below a series of related contents: links to articles, videos, related apps and results of previous research.
For some objects, the research is surprising: recognizing a keyboard the least, but understanding that a "handmade" bag, not immediate. In other cases, it happens that a tape recorder is mistaken for a planer, but then tries with usb (already closer), "device", "equipment".
On the homepage the service offers a series of results to be navigated, already built. They range from "Game of Thrones" to "Hillary Clinton", passing through UefaEuro 2016.
By trying it, you understand the point of view of the CEO, Ambarish Mitra: speaking with Business Insider he said that he would not sell to Google, because what he works on "bigger than the Internet itself".