Automated Filing of Scanned Documents

From $1

1. Creating the Rules
1. 1.1. ID
2. 1.2. Patient Name
2. Given
1. 2.1. NHI/SSN/DOB
2. 2.2. Test Date

Many documents received in a medical office are very similar, and based upon this similarity, smart scanning can determine for whom a scanned image refers. The scanned image is examined at certain positions according to a special definition which the user creates. The text is taken from each position and if there is enough information, and the text high enough quality, the image can be filed with a high degree of certainty that we have associated it correctly.

Creating the Rules

In the Settings/OCR/Autofile tab:

click on [Load Scan] - you need to pick a representative PDF ( or TIFF/g3) as this will be used for the basis of subsequent recognition attempts
The PDF is converted by ghostscript to a PNG ( smaller size to make it fit on screen )
A paint type program then loads up the PNG.
You now need to use each to the buttons below the [Clear] button to create the definition

ID - use your mouse to select a piece of text that uniquely identifies the form. Click on the [ID] button to transfer the coordinates of the box to the form. A larger version appears so that you can confirm that you have not picked up any other characters. You will have to wait while the OCR webservice recognizes the text. This then appears at the top of the window.
Patient Name - repeat the same for the region where the patient name appears. If the surname and first name are on different parts of the form, just use the surname region.
Given - use this field only if the given name appears on a different part of the form from the surname
SSN/NHI/DOB - and for the patient identifier
Test Date - and for the date on which the test was performed.

ID

You now need to specify what the text is that you selected with the [ID] button. Synapse will try to identify this form by looking at the specified region and to see if the text there matches what you specify.

Patient Name

The patient name format has to specified as well. This is normally [surname first-name middle-name]. If the name moves around a little on the form, but is always preceded by some text like "Name: " you can put this in the Prefix field. You can add some fixed text to the end of the format such as "dob: " like [ first-name middle-name surname dob: ] which is helpful if all the details are running along one line.

Given

If the surname and given names are on different parts of the form, then use this field to get the first of the given names. The format should be "first-name".

NHI/SSN/DOB

A drop list is available to specify the format for the patient identifier. These are:

alpha - alphabet characters only
numeric - numbers only
alphanumeric - mixture of numbers and alphabet
date - numbers and "/" character
anything - no constraint

Test Date

The test date format also needs to be stated. It might be [ dd MMM yy] for [21 Nov 07], or [ mm dd yyyy ] for [ 11 21 2007 ], or [ mm dd yyyy] for [ 11/21/2007 ]. Again you can add a prefix if the date position shifts, and you have some prefix text available. Note that each of the day, month, and year elements in your date format must be separated by a space. Synapse will strip out any "/" or "-" from the ocr'd test date.

The possible formats are:

dd = a 1 - 2 digit day
mm = a 1 -2 digit month
mmm = a 3 character month
month = the month spelled in full
yy = a 2 digit year
yyyy = a 4 digit year

Note that Prefix text must be included in the region you specify for scanning.

Now save the rule, making sure that you have provided a rule name.

Retrieved from "http://compkarori.no-ip.biz:8090/Documentation_Index/Automated_Filing_of_Scanned_Documents"

Tags:

Edit tags

Attach file or image

Files (0)

Images (0)

Comments (0)

You must login to post a comment.