You have successfully implemented Optical Character Recognition (OCR) in your DLP environment
However in the Enforce console, you notice a significant number of alerts with the code 4803 - OCR Request was not successful
Clicking on the details of the alerts shows the subcategory: Image Size Error or Bits Per Pixel error
For Example, you can find the coinciding errors in the ocrworker-OCRWORKER log from the OCR server.
Error while loading page: Image error: [8004c70e]: IMG_SIZE_ERR - *2363063181216,3924 (0)
Received OCR request ID [0e2159ef-b05f-4a58-8e9e-9ff369d681b8], image size [620] bytes
Another example, regardless of the size of the image, if the color resolution (Bits Per Pixel) is too low, grayscale, for example, dlp may not be able to distinguish any characters.
03/13/23 12:10:00 | WARN | ocrEngine | NuanceEngine | Error while loading page: Image error: [8004c708]: IMG_BITSPERPIXEL_ERR - *2618016796576,329332 (0)
When using OCR with Email Prevent, many times the images in Signatures are too small or not enough color resolution for detection and will cause these errors.
Release : 15.x
Component : OCR
These alerts are because the OCR pre-filter on the detection server is passing images to the OCR servers that have sizes or color resolutions that are unsupported, either too big, too small, or too fuzzy and are unsuitable for OCR extraction.
The OCR server then drops that file and posts the alert to the Enforce.
Alerts of this type can be safely ignored, no extractable data was available in those images, so OCR detection is not failing.
Refer to article Image Quality and Resolution for OCR results for image specifications for OCR results.