Select Fast if you have a high volume of images (the load level on your OCR server will be large), and are concerned about performance. Only large, text-intensive images are sent for extraction; small images and documents that don't contain much text are not extracted at all. This option enhances performance, but may sacrifice accuracy.

Select Accurate if you have a small number of images (the load level on your OCR server will be small). Every textual image in your network is sent to the server for extraction. This affects performance, but provides the most accurate results. If response is inadequate—for example, browsers are timing out on the HTTP channel—change this setting to Fast or Balanced.

Select Balanced (default) for a balance between accuracy and speed.