A little while ago Google released an update to Google Translate that greatly improved is ability to recognize and translate text within images, and to overlay the translated version of the text into the image in near real-time. This is certainly something worth playing with to see how well it holds up to real world usage. Surprisingly I encounter a fair amount of signs and documents that aren’t online and also aren’t written in English.
There is a colorful poster on Etsy written in French that I used as my test for how well this works with a simpler set of words. I may actually pick this up for my younger cousins, it really is a nice poster.
For those not fluent and don’t recognize words similar to their English translation these are the colors yellow, orange, red, pink, purple, blue, green, brown, and grey. I am not fluent in French and had to look up purple, that one never sticks with me for some reason.
To get to the App to the point I first needed to install the language I was interested in having available offline. I actually wound up downloading a few languages since I tend to encounter a few in my travels around Manhattan. And if this works as described I can see using this when I travel later this year to Germany.
After the install it was very easy to get the camera turned on and the translation to start. Though it was easy that does not mean it is highly accurate. In this example I saw the text popping in and out, trying to translate the words but not quite able to recognize each line as a singular word. This caused some odd translation attempts to occur.
What I found to be the most reliable was to use the refinement tool and allow it to handle the translation without the real-time overlay on the image. In this scenario the translation came back quite accurately.
This time the translated text is much more on point and gets all of the colors properly translated, but not over the poster image as I was hoping. Not quite perfect, but this is a fantastic advance in real time text translation and I look forward to using this in my travels. Though not making the most of all the features in the App I can see myself using this the most since after a few trials this really gives the best results.