Autofile

Requirements

Getting started wth Autofile

The first thing you need to do is create some rules for each type of document.  I suggest that you start with only a couple of document types so that you understand how it works.  A Synapse tutorial is here and much the same applies.

Your rules are saved in a file called rules.r - do not delete this as then you will have to redefine all your rules again.  You should also keep a copy of the files you used to create the rules in case you need to recreate them, or to check that the regions are correct.

Once you have defined some rules, you then select the directory where your scans are, and then click on the [RUN] button.  Each file will then be processed.  You may have to tweak your rules if it is unable to recognize documents for which you have defined a rule.  You may have to create a larger recognition area or reduce it if other text is being captured.  If the file is from a fax you may need to increase the vertical area of your box to account for changes in vertical displacement.

When the debug check is ticked, you will see some information on what Autofile is seeing which may give you some clues.

Because of the way the images are converted from PDF and TIFF to PNG, the sizes are not the same for both.  So, you can not apply a rule derived from a TIFF image to a PDF, and vice versa. 

Tutorial

Here's a Jing video tutorial on how to create a rule

NB: There is a current bug that can lose some of your rules if you update a rule after you do the recognition.  Best to update when you first start off Autofile for the moment. 

Release History

• Build 23: Removes any spaces found in the NHI zone.
  
• Build 22: Crops the image horizontally as welll as vertically.
  
• Build 21: If a rule is selected in the rules table, it will be used as the first rule.  This is where you know which document type you are importing.
  
• Build 20: Switches back to Web OCR if get non alpha characters in patient's name
  
• Build 19: If using Tesseract, and OCR fails on the date, it switches back to the Web OCR to try again.
  
• Build 18: Option to use Tesseract on high quality scans.  Regions are now cropped before OCR.  Files are copied from the source directory and renamed prior to OCR to fix paperport filenames which were too long
  
• Build 17: Adds support for g3 raw fax format
  
• Build 16: Now has an [Install] button to download all the necessary files
  
• Build 15: brackets the zone for scanning
  
• Build 13: Bug fix for date errors - not binding to date rule
  
• Build 12: Bug fix for build 11, was not moving the files
  
• Build 11: If debug is ticked, then the tesseract ocr'd text is written to the directory ./tessoutput
  
• Build 10: More network error checking, and now waits if service is temporarily unreachable
  
• Build 8: Does a scout scan using tesseract which improves speed considerably
  
• Build 7: Optional given name field, and larger image screen for selecting regions
  
• Build 5: Adds TIFF support
  
• Build 4: Adds a [Load Coords] button so that definitions can be displayed on a PDF
  
• Build 3: Removed residual debugging code still present