Google looks to patent technology that recognizes text in images

Computerworld - Google Inc. may have plans to expand its search engine beyond static text.

The search giant in June filed a patent application for technology that can recognize text in images. It could be used to retrieve text from video or from photographs that may show up as part of a street scene. The application was published on Thursday.

The company is seeking the patent for methods, systems and computer programs that can extract image text. For example, the application noted, the technology could search a collection of keywords extracted from text and retrieve an image associated with the extracted text.

The digital images that the technology aims to scout for text include illustrations of landscapes, people, urban scenes and other objects, according to the application.

"Image text typically includes text of varying size orientation and typeface," the application noted. "Text in a digital image derived, for example, from an urban scene (a city street scene) often proves information about the displayed scene or location. A typical street scene includes, for example, text as part of street signs, building names, address numbers and window signs."

Duncan Riley, a blogger at TechCrunch, noted that if Google really has found a way to index text in static images and video, "this is a great leap forward in the progression of search technology. This will make every book in the Google Books database really searchable, with the next step being YouTube, Flickr and more."

Adam Ostrow, a blogger at social networking site Mashable, said  that potential applications of text recognition technology go beyond Google's existing products.

"Google could start photographing the aisles of stores to create product indexes with vast amounts of data," he wrote in a blog post. "For example, if Google were to photograph a box of Lucky Charms, it might be able to tell you not only its price, but nutritional information and ingredients."

However, Ostrow noted that the amount of manpower needed to capture this type of data would be hefty.

"Google Street View provides a fair amount of photos for major metropolitan areas, but it would seem that creating meaningful search products that utilize text recognition technology would require far deeper and more precise imagery," he said. "It sounds challenging, but perhaps not impossible with 15,000-plus employees at the company's disposal."

Read more about web 2.0 in Computerworld's Web 2.0 Knowledge Center.