Optical Character Recognition (OCR) recognizes text within images, enabling you to gain further clarity on images added to your Check workspace.

This article explains:

How OCR works

Within each image item, you can view a snippet of the extracted text in the media panel. In case the full text is hidden, click More to display the full extracted text.

How to manually extract text from images

For manually added images, from the Image actions menu on the media panel, click Image text extraction.

Similarity between OCR and other medias

Once text extraction is done, text items that are similar to the image will be suggested.

When an image is added to Check, the text is extracted. Later if an audio or video is added to the workspace, and it contains a text, that will be extracted, and this text is similar to an existing image on the workspace, the new text item will then be suggested as a match to the image item.

Learn more about similarity matching and suggestions.

Extracted text search

There are two ways to search for image items using extracted text:

Regular search:

  1. In the search bar, click the gear icon to the right, select Extracted text (OCR), then click the Update blue button.

  2. Input the text in the search box to display items matching it.

Similar item matching:

  1. Open an item, and in the media panel click Add similar and select the Import similar media from other items option from the drop-down list.

  2. Paste (extracted) text into the search to reveal image items that match.

OCR supported languages

Currently, we support text extraction in these languages: Arabic; Bengali; English; Filipino; French; German; Hindi; Indonesian; Malayalam; Marathi; Portuguese; Spanish, and Telugu.

Notes: We don't extract text from the media within tweets or Facebook posts. We extract the text directly added to the links. If you want to extract the text from within a media, directly add the media to the workspace, by either using the image link or adding it manually.

Did this answer your question?