Module Description - Instructional Outcomes - Instructional Events
Instructional Evaluation - Discussion List

Section 7: Correcting the OCR

When Kurzweil scans a page it displays a picture of the page on your desktop. However, Kurzweil also runs a second operation on the text called optical character recognition, or OCR, which converts that picture to readable text.

When Kurzweil reads this page we hear errors on the words, “between” and “civilization.”
(Kurzweil reads the highlighted text on the screen, mispronouncing “between” and “civilization.”)

Apparently the “t” in the word “between” has been interpreted as an “r”, which is understandable given the font. The “c” in “civilization” has been recognized as a separate unit, which is not unusual when these large capital letters are used at the beginning of a passage.

Let’s fix those errors by choosing Edit -> Correct OCR-> Edit Underlying Text.

Here we see four lines of text. On the top of each line you see the picture of the text as it appears on your screen. Underneath each line of picture text you see how Kurzweil interpreted the text. This is the text the computer voice will read. This is also the text that will be exported if you save the document in another format, such as Microsoft Word.

To edit the text, click on the word and make the changes in the edit box below.

We’ll just erase the “c” here and add it at the beginning of “civilization.”

Click the Read button to test your correction. (Kurzweil reads, “civilization.”)

You can use the Next and Previous buttons to navigate within the document.

When you have finished making your changes, click OK.

When you return to the document, Kurzweil will read the corrected underlying text.
(Kurzweil correctly reads, “Land Between Rivers” “Civilization develops only when people have plenty of food, and only farmers can produce plenty of food.”)

This concludes Section 7.