Google has launched an updated version of its photo-based instant translation Smartphone application.
The Google Goggles prototype was revealed in February, and was able to translate text from German to English. Users would point their phone at the text on a sign or menu, for example, and Google would immediately translate.
Now the application has been widened to be able to render French, Italian and Spanish into English.
Version 1.1 also features "improved barcode recognition, a larger corpus of artwork, recognition of many more products and logos, an improved user interface, and the ability to initiate visual searches using images in your phone's photo gallery," according to a Google blog post earlier this month.
Google Goggles so far only supports Latin-based alphabets, but developers say they are working on improvements and "hope to support Simplified and Traditional Chinese in the future."
Google software engineers Alessandro Bissacco and Avi Flamholz acknowledge that "computer vision is a hard problem," and product reviews have noted the limitations.
Tech journalists at the San Francisco Chronicle followed a Google spokesman through the San Francisco Museum of Modern Art, and observed him showing "how the tool worked – or in one case didn't – on paintings, barcodes, foreign-language newspapers and more."
Ced Kurtz at the Pittsburgh Post-Gazette said Google Goggles was "cool, but … still a little foggy."
Image recognition is a major challenge for developers, and automated, computer-based translation is a process that still falls short of what human translators can accomplish.
16 May 2010
Go back to May 2010
