Pick Region dialog (Tesseract OCR Engine)

Overview

The OCR action provides the option to select a specific region within an image for more precise text retrieval using the Pick Region dialog.

To access the Pick Region Box for the Tesseract OCR Engine

  1. From the Task Builder, expand the OCR action, and then select the desired OCR activity.
  2. Click the OCR Engine box, and then select Tesseract.

  3. Select the Specified region (improves accuracy) parameter, and then click Pick Region.

To pick a region using the Tesseract OCR Engine

  1. Click the Languages box, and then select the languages that correspond with the text you want to retrieve from the image.
  2. For the Get text activity, select Exact copy (do not format text) if you want to retrieve the text as it is formatted in the image (optional).
  3. In the Image Preview pane, hold down the left mouse button, and then drag the mouse pointer to draw a box around the region you want to retrieve text from.
  4. The OCR Preview pane updates and displays a preview of the text retrieved from the specified region. If the desired text was not retrieved, repeat step 3, or click the Page Segmentation Mode box to select a different mode to attempt a more accurate scan, based on the position of the words to retrieve. The available modes are:
    1. Auto Osd - Provides automatic page segmentation with orientation and script detection (OSD).
    2. Single Column - Assumes the lines of text in the image are in a single column of text, varying in size.
    3. Single Block Vertical Text - Assumes the lines of text in the image are in a single uniform block of vertically aligned text.
    4. Single Block (default) - Assumes the lines of text in the image are in a single uniform block of text.
    5. Single Line - Assumes the entire image only contains a single line of text.
    6. Single Word - Assumes the image only contains a single word.
    7. Circle Word - Assumes the image only contains a single word contained within a circle.
    8. Single Char - Assumes the image only contains a single character.
    9. Sparse Text - Scans the entire image to retrieve as much text as possible in no particular order.
    10. Sparse Text Osd - Scans the entire image to retrieve as much text as possible using orientation and script detection (OSD).
    11. Raw Line - Assumes the image only contains a single line of text, bypassing hacks that are Tesseract-specific.
  5. To edit the region without using your mouse, manually adjust the Top, Left, Width, and Height pixel coordinates, and then click Refresh to update the Image Preview pane.
  6. If you select or clear selections for the Languages or Exact copy settings, or change the page in a multipage file after picking a region, click Refresh to update the Image Preview pane.
  7. When you are finished, click OK to set the region.