Recognize Text in a Single Document - Best Settings

Question

Scanning and OCR Question:
We are currently using Adobe Acrobat Professional Version 9.  The scanner is a Canon DR-9080C.  What are the best setting to assure the best searchable image?  Once we upload the pdf document into our database we want to make sure it is searchable and selectable.  Also what is the best scanning dpi setting?
Currently when we OCR a document and insert into our database we are unable to get accurate search information.
Any help would be appreciated.
Thank you,
[email address deleted by host]

Answer 1

The OCR output of Searchable Image and Searchable Image (Exact) uses text rendering mode 3.
Text rendering mode 3 renders glyph outlines with neither fill nor stroke (invisible/hidden text).
Use a Preflight profile that evaluates for "Invisible text objects".
This finds text objects which use text rendering mode 3 (invisible text).
Create a new Preflight profile and place it in the "Custom profiles".
Give the profile a friendly name (perhaps "Invisible text").
Select Custom checks.
In the 'find' field [1] type "inv" then select "Invisible text objects".
Click the arrow [2] to include this selection in the profile [3].
capture 1
While the default indicator is "Error" for when a Preflight fires you can change this to Warning or Info. [1]
To further configure the profile click the "Edit check" button. [2]
The default trigger for this custom check is "is true" but if you want "is not true" you can change that in the Edit Check
dialog.
capture 2
Once you have this (or any profile created) Preflight profile saved export it to a dedicated location (say a "KFP" folder).
When many PDFs, in a common folder, are to be assessed create a Batch Sequence (pre-Acrobat X) or an Action
(Acrobat X) that uses the Preflight profile.
Using Acrobat 9 Pro as an example --
Create a Batch Sequence. One I use is "Preflight HT".
Click Edit Sequence...
capture 3
This opens the dialog "Edit Batch Sequence - Preflight HT" . Click the Select Commands button.
capture 4
Next, select "Preflight" then click the "Edit" button.
capture 5
In the dialog that is presented you can configure reporting options.
Ticking "Display a summary PDF with a list of files for which Preflight found errors or warnings" will provide a PDF report
that facilitates a quick at-a-glance means of assessment.
Once configured click the "Save" button.
capture 6
Next you'd click OK on the Edit Sequence dialog to close it.
In the "Edit Batch Sequence - Preflight HT" dialog you'd click on the "Output Options..." button.
In the Output Options dialog configure configure as desired.
capture 7
The summary report provides useful information.
An icon adjacent to the file name provides a visual indicator. [1]
For an Information alert you have blue. Note that "information" is denoted as "Notification".
For a Warning alert you have the yellow/orange icon.
For an Error alert you have the red icon.
For each file you are provided with a count of error(s), warning(s), and notification(s).
The file name in the summary report is linked to the file which lets you view the file for specifics.
capture 8
What can be useful is to create an Action / Batch Sequence that use multiple Preflight profiles.
As an example is to check for fonts not embedded, invisible text present, and text can be searched.
The premise is that PDFs in a given folder all contain PDF page content that is scanned images of textual content and that
all PDFs have undergone OCR (Searchable Image or Searchable Image (Exact)).
The output summary report can provide V&V that the premise is true.
--| OCR output (text rendering mode 3) is, by default, not embedded
--| The OCR output is expected to be searchable.
--| OCR output *is* of text rendering mode 3; so we expect this to be present.
Be well...