The Google Lens translator does more magic with this novelty that will arrive at the end of the year on your mobile

Google’s multi-search or the new immersive view of interiors Google Street View have not been the only novelties that Google has presented during its Search On event. The company has also announced a major improvement for Google Lens translator.

Among all the functions of Google Lens we find one of the most used: its augmented reality translator in which translates more than 1,000 million images every month and that in a few months that translation will be much better. And it is not because it is going to offer better ones, something that Google works every day to improve them, but rather translations will be visually much better.

Translation perfectly integrated into your image

What you love about Translating with Lens is now even better. 💡

With major advancements in AI, translated text appears seamlessly integrated, as if it was part of the original picture. Turning text… into context! #SearchOn pic.twitter.com/N8YySv87z1

—Google (@Google) September 28, 2022

Currently if you focus on a text with the Google Lens translator camera, or select an image to translate it, the application overlays the translation on top of the original text with colored blocks, so if an image has a lot of text during the translation, a significant percentage of that image will be lost. Well, that won’t be the case in the near future.

Google has announced that thanks to great advances in AI, the translated text will appear perfectly integrated, as if it were part of the original image. Once an image has been translated, it will be very difficult to know what language the original was in.

To achieve this, what Google Lens is going to do soon is separate the text from the image. Image recognition will translate the text and put it in the most similar font over the image, but before pasting the translated text, Google’s magic eraser will erase the original text from the image so as not to have to put colored blocks to hide them.

This image and text processing will be done in 100 milliseconds. In the duration of a blink we will see the translated text with the background of the original image, either by translating an image or in real time by pointing the camera at a poster. Google has reported that this novelty will arrive at the end of the year.