|
 |
 |
|
| |
Whether to OCR, or not?
There are really two moments when you should think about OCR ’ing documents. The first moment is ahead of time, in general. The second moment is when you are thinking and planning about a particular document production. We will only deal with the first question here.
OCR ’ing, in general, can provide speedy access to a stack of documents, but it is not accurate enough for litigation. The larger document firms estimate an 80% accuracy rate for legal documents. Some of the fancier firms and programs can OCR as they scan and will find many, if not all, names in the text of the document and insert those names into a Name field in the database, but those kinds of tools are seldomly used in small lawsuits.
If you are handling a small lawsuit, and all you want to do is keep track of the documents, highlight the good stuff so that it can be found again, and send information on to where it can be used for further discovery and trial preparation, you may be tempted to OCR the documents as part of processing them. Bear in mind, however, that OCR ’ing a PDF file with the Capture plug will change the image’s appearance. See below. The first example is an order that was scanned in as an image file. That image was then OCR’ed and the result is the second example. The darker text in the second example resulted from the recognition of only some of the words in the text. The differences are even more apparent when the document is printed out. In short, if you OCR a scanned image, you will change its appearance.
© Copyright 2002 Ernest Svenson.
Last update: 3/9/2002; 3:20:34 PM.
|
|
|
