Google Adds Text Recognition To Goggles
The original Goggles was designed to let users of devices running on Google’s Android operating system take a picture of something – such as a landmark building or a product’s barcode – and find information about that item.
The new Goggles v.1.1 extends that capability to text in images, giving it the ability to read text in five languages: English, French, Italian, German and Spanish. Google is working on extending the feature to non-Latin languages such as Chinese, Hindi and Arabic.
Google first demonstrated the Goggles translation capability earlier this year at the Mobile World Congress in Barcelona, but at the time it could only recognize German text.
“Computer vision is a hard problem. While we are excited about Goggles v1.1, we know that there are many images that we cannot yet recognize. The Google Goggles team is working on solving the technical challenges required to make computers see,” wrote Google software engineers Alessandro Bissacco and Avi Flamholz in a blog posted Thursday.
With the feature, users point their phone at a word or phrase and use the “region of interest” button to draw a box around specific words, and press the shutter button. If Goggles recognizes the text, it will give the user the option to translate it and select the source and destination languages.
In addition to the new translation capability, Goggles v.1.1 offers an improved user interface, enhanced barcode recognition, a larger database of artwork, the ability to recognize more products and logos, and the ability to initiate visual searches using images in a phone’s photo gallery.
Google Goggles v.1.1 is available to run on devices running Android 1.6 and higher.