Explained: How Google Lens image text copy feature works

If you’ve used Google Lens then you must be aware that you can simply extract texts from an image using the app and paste it in a message, make calls and even add the extracted text in a document. But, have you ever wondered how the entire conversion from image to text happens? How does the Google Lens app identify texts in the image, font and language in the image. Let us explain.
Ana Manasovska, a Zurich-based software engineer who has worked with Google Lens’ ability to convert an image to text. According to the official blog post, Manasovska has explained that she has been involved with the text aspect for the app to make it discern text and copy it for a search or translate it.
How does Google Lens manage to identify text, language and structure of text from an image?
Manasovska mentioned that Google Lens uses Machine Learning which they like to call as models which are trained by her team to identify characters and structure in images. The machine learning model in Google Lens has the ability to teach itself to distinguish text structures just like a human would.
Also, the machine learning model in Google Lens is trained to recognse 30 scrips, including Cyrillic, Devanagari, Chinese and Arabic. Manasovska has mentioned that the model is most accurate with Latin-alphabet languages at the moment
The entire programming for the app has been done on C++ as it allows developers to run processing steps needed.
Google Lens features
For those unaware, Google Lens comes in handy for lots of things. The main feature here is the identification of the texts from an image and make it accessible for copy, paste, translation and more. In addition, users can also use the Google Lens app to run a search from an image. All these features, according to the official blog post, is based on the machine learning capabilities and also the well-trained models.



Read original article here

Denial of responsibility! TechnoCodex is an automatic aggregator of the all world’s media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials, please contact us by email – [email protected]. The content will be deleted within 24 hours.

Leave a comment