How to Get Live Text OCR on Photos Right Now from iOS 15
One of the innovations in iOS 15, which Apple unveiled at the developer conference on June 7, was the Live Text feature. Its purpose is to read the text in photographs taken with a camera. This can be a snapshot of a board, document, receipt, label, and much more. All you need is to click on the recognition button, and then select and copy the resulting text. The thing is definitely very cool and comfortable. But you don’t have to wait until iOS 15 is released to take advantage of it.
As you can probably imagine, the Live Text feature is not exclusive to Apple. Because applications and services that can recognize text in photos and copy it has been around for a very long time. So the choice will obviously be large, do not hesitate.
How to copy text from a photo
Most of these applications, as a rule, are either independent solutions that can no longer do anything or, on the contrary, can do too much, which frightens users. So it’s not very convenient to use them. Fortunately, Google Photos is a welcome exception to this rule.
Yes, yes, Google Photos also has text recognition on photos. Here’s how it works:
- Download the Google Photos app from the App Store ;
- Run it and upload a photo with text there;
- Click “Copy text from image”;
- Select the desired fragment and click “Copy”.
Typically, when you open a photo in the Google Photos interface, the application algorithms immediately determine that it contains the text. Therefore, the offer to copy it most often appears automatically. But, if this did not happen, click on the “Google Lens” button (second from the right).
Thus, you can recognize not only printed text on paper but also on other media. For example, Google Photos’ algorithms do a great job with signs, pointers, banners, and even handwritten text. True, it is desirable that the text be written in block letters because the application has quite big problems with writing.
Recognizing text in photos
But Google Photos can not only recognize text but also offers many options for subsequent interaction with the result. For example, you can translate it into another language, search for information on it in Google, listen to it, and also transfer it to your computer when the synchronization function is enabled.
In fact, Google’s recognition algorithms work well enough that there are usually no mistakes. However, the recognition accuracy depends on the quality of the typing of the text, the lighting, and the angle at which the picture was taken. See how many mistakes the application makes on the example of cheese packaging and how accurately it recognizes everything on the Big Taste box.
Poor recognition quality in the case of a cheese label is due to bruising. Because of this, the entire text is not in the same plane, but is, as it were, divided into fragments, which, moreover, are quite noticeably displaced relative to each other. Algorithms have to work hard to recognize everything, and this provokes them to make mistakes. In the case of “Big Tasty”, the field for recognition is clearly better.