Optical character recognition (OCR) support

FortiProxy 7.4.0 adds support for optical character recognition (OCR), which detects and extracts text in images with a score greater than the configured OCR threshold and takes action accordingly as defined in the DLP profile.

Under Content Analyses > Image Analysis, when creating or editing an image analysis profile, use the new Optical Character Recognition option to enable or disable optical character recognition. When enabled, you can further configure the threshold to activate optical character recognition. The value range is 0 - 100, where 0 means bypassing all images without DLP and 100 means processing all images with DLP that triggers OCR. The default is 30.

The Optical Character Recognition option takes effect only if the policy meets the following requirements: Content Analysis is enabled with this image analysis profile selected. DLP Profile is enabled with a profile configured and selected.

To configure OCR in the CLI:

config image-analyzer profile
  edit "test"
    set porn-status allow
    set image-skip-size 100
    set image-skip-width 100
    set image-skip-height 100
    set optical-character-recognition [enable | disable]
    set ocr-activation-threshold [integer]
  next
end

Refer to config image-analyzer profile in the CLI guide for more information about using this command.