Skip to main content

Screenshot / Image Translation

This feature can automatically translate screenshots in the clipboard with global shortcut keys, for details, please refer to Translate Content Outside the Browser.

Screenshot Translation​

划词翻译 supports capturing screenshots from web pages and translating them. The use cases are as follows:

There are several ways to open screenshot translation:

  • Right-click anywhere on the web page and choose 划词翻译 > Screenshot Translation.
  • After opening the translation panel, click the scissors icon in the upper left corner.
  • After configuring the shortcut keys on the settings page, use the shortcut keys to open.

After opening, the mouse pointer will become a cross, and you can take a screenshot on the web page. After taking the screenshot, click the OK button to translate.

Image Translation​

In addition to using 划词翻译's screenshot translation to take screenshots on web pages, you can also translate image files on your computer or read screenshots in the clipboard (for example, after taking a screenshot with WeChat's Alt + A or Windows' built-in Win + Shift + S).

After opening the translation panel, click the landscape icon in the upper left corner. In the pop-up window, you can choose to select an image from your computer or read an image from the clipboard. After selecting, click the OK button.

Why is Screenshot / Image Translation Slow?​

Screenshot / image translation will first convert the image to text, and then send it to the translation panel for translation. By default, 划词翻译 uses the built-in Tesseract OCR to convert images to text, which can be used without any configuration, but has the following disadvantages:

  • It does not support automatically detecting the language in the image, so when you use screenshot / image translation, you need to manually select the language of the image.
  • Whenever it translates a language for the first time, it needs to download a detection package of about 10MB, which will cause a long waiting time.
  • Even if the detection package has been downloaded, depending on the size of the image, it also takes 2 ~ 8 seconds to convert the image to text.
  • Its accuracy is not very high.

In view of this, 划词翻译 also supports using the Baidu General Text Recognition service to convert images to text, which does not have the above disadvantages. The recognition only takes less than 1 second and has a high accuracy, but you need to apply for it yourself, but the whole process will not exceed 3 minutes. Please read Apply for Baidu General Text Recognition Service for application steps.