DLP - OCR

Use text recognition on images (OCR)
If this option is enabled, then during DLP operations, if there are individual images or images inside document files, the image will be converted into text for subsequent text DLP analysis.
Basic DLP settings on the client are made on this page, and the OCR process itself occurs on the server, settings here.

Maximum wait time per document in seconds
It is important to understand that the OCR process can take a significant amount of time, so sometimes it makes sense to set a maximum wait time (for the entire document).
For example, if the server is under heavy load and the timeout expires, the OCR result will be returned as "OCR timeout", which you can optionally use in DLP sensitivity lists. The user will then be able to resend the file (if sending blocking is configured).
If sensitivity to the phrase "OCR timeout" is not set, then in the case of a timeout, this will simply mean that not all images in a document with graphic images will be recognized and text extraction will be incomplete.
To display the document analysis process to the user, a small window will appear on the screen. Settings are configured on this page.

Minimum and maximum image size in bytes
For optimization, you can specify a range of image sizes for popular graphic formats in bytes, so as not to waste OCR time on obviously "useless" images.

Enable caching on clients
This saves time by eliminating the need to send previously successfully OCR-ed images to the server. This is especially useful when an employee submits the same document multiple times. The images themselves are not cached on the client, only their short hashes are stored.