Recognition Engines

This step configures the type of automatic recognition performed as images are captured. Set options in this step for Barcodes, Patch Codes, Blank Pages, Characters (OCR/ICR), and Marks (OMR).


Barcodes

When separating documents using barcodes or populating index fields with barcode values, the user must select the ‘Enable Barcode Detection’ box. Once Barcode Detection is enabled, the program will attempt to recognize any barcodes during the capture process.

Barcode Engine - The two engines available are Standard Engine and Legacy Engine. The following options are available once Barcode Detection is enabled:

Barcode Types to Detect

There are many types of Barcodes that can be used on documents. Select the barcode type or types the user wishes to recognize for this Document Type. NOTE: Selecting all the possible barcode types listed and not just the barcode types that are not necessary will cause a potentially significant loss of performance during capture.

1D Barcodes

The following 1D barcode font types are supported:

Australia Post, Codabar, Code 11, Code 128, Code 32, Code 39, Code 93, EAN-13, EAN-8, Intelligent Mail, Interleaved 2 of 5, ITF 14, Planet, Plus 2, Plus 5, Postnet, RM4SCC, Rss14, Rss 14 Limited, Telepen, UPC-A, UPC-E

Press “All” to select all barcodes. Press “Clear” to clear.

2D Barcodes

The following 2D barcode font types are supported: Data Matrix, Micro QR, PDF417 and QR

Preview

 - Load a sample page containing barcodes to determine the types of barcodes present and check for functionality.

Options

Chose “Selected Barcode Types Only” if the user has already determined specific types of barcodes that will be in use. Chose “All Supported Barcode Types” if the type of barcode is unknown.

Recognition Results

 - Initiate detection. Each barcode will be listed in order of appearance on the page next to their type and values.

Recognition Engine Options

Maximum Barcodes per Page

The user may further optimize the barcode recognition accuracy and performance by specifying the maximum number of barcodes expected on any one page. The barcode recognition engine will stop processing subsequent barcodes once it has reached the specified number of barcodes per page.

Scan Interval

Additional optimization can be achieved by specifying the scan interval that the barcode recognition engine uses when looking for barcodes. The scan interval is the number of pixel rows it should skip when checking for barcodes. So when the value is set to 10, the engine will check every 10 rows for a barcode. Increasing the Scan Interval, will increase the speed of detection process, but at the same time can lead to omissions. Conversely, lowering the Scan Interval will slow barcode detection, but will result in more comprehensive detection. The default level of 10 as it balances both the speed of the engine while allowing a high level of detection. This interval is DPI dependent as a 200 DPI 8.5 x 11 page has 2200 pixel rows and a 300 DPI 8.5 x 11 page has 3300 pixel rows.

Barcode Orientation

If barcodes will be captured either horizontally or vertically, select the appropriate option to enhance the speed of barcode recognition. If unsure of the direction by which the barcodes will be captured, select all of the directions to ensure that all barcodes will be read correctly. If “Support mis-formatted 3 of 9 barcodes is selected only “Left to Right” and “Top to Bottom” are available.

Other Options

Enable Enhanced Barcodes

Selecting this feature enable a Code 39 to support its extended character set including + which can change the character following it to be read as lower case instead of the default which is upper case only.

Support mis-formatted 3 of 9 barcodes - A mis-formatted 3 of 9 barcode (Code 39) is one whose first 4 bars are NOT (skinny, skinny, fat, fat) and/or final 4 bars are NOT (skinny, fat, fat, skinny). Setting this option could output all kinds of unexpected data when reading correctly formatted barcodes.

Truncate Whitespace

Selecting this option will cause all spaces at the end of the barcode to be removed once the value is placed in an index field.

Skip recognition on back pages

Selecting this option will skip barcode detection on back pages, leading to increased performance during capture.

Barcode Splitting

Split Characters

Should the desired barcode contain multiple data values separated by a character, enter said character here in order to split the single barcode into multiple values. If the desired barcode contains 111,aaa,222enter , in the Split Characters box. For purposes of index the program will automatically parse or split the barcode into 3 logical barcodes, the first barcode will contain 111, the second will contain aaa and the third will contain 222. Further parsing options for barcodes based on fixed position are available in Step 5 of 9 (Index Data Fields) in the barcode tab.to split on tab use <TAB> and to split on new line use <NEWLINE>.

Zones to Recognize

User can chose to detect the entire page (default) or define a zone for the recognition to process, further increasing performance. Hit “Define Zones” to proceed with Zone Configuration.

Zoom in to the desired area leaving room to maneuver. Then select the draw zones Icon and draw the zone(s). Name the zone and fill out which page of the document it is expected to be found on. The Zone Names are kept in a list for use anywhere in the program that the user can Define Zones. NOTE: The Page of the template and its Resolution is displayed at the bottom of the screen and they MUST match the page and resolution at capture time.

 - Delete the highlighted unwanted zone. Note: If a Zone is in use by any Document Type, the zone cannot be deleted.

 - Option unavailable from this zone configuration window. Used for grouped zones.


Patch Codes

Enable Patch Code Detection

Select this option to cause detection of patch codes and the program will attempt to recognize all patch code types (I, II, III, IV, VI, or T) during the capture process. NOTE: Setting what to do after detection is defined in the Document Type Configuration/Separation Options section of this manual.

Recognition Engine Options/Reading Quality

Select between most accurate or fastest.

Patch Code Orientation

Define which page orientations the recognition engine will attempt to recognize patch codes. Select the appropriate direction(s) to enhance the speed of patch code recognition. The default is all directions and it is recommended that the user choose the default.

  • Left to Right
  • Right to Left
  • Top to Bottom
  • Bottom to Top

Zones to Recognize

Define which zones (default is entire page) to attempt patch code recognition in. Any patch codes outside the defined and selected zones will not be recognized. 

User can chose to detect the entire page (default) or define a zone for the recognition to process, further increasing performance. Hit “Define Zones” to proceed with Zone Configuration.

Zoom in to the desired area leaving room to maneuver. Then select the draw zones Icon and draw the zone(s). Name the zone and fill out which page of the document it is expected to be found on. The Zone Names are kept in a list for use anywhere in the program that the user can Define Zones. NOTE: The Page of the template and its Resolution is displayed at the bottom of the screen and they MUST match the page and resolution at capture time.

 - delete the highlighted unwanted zone. Note: If a Zone is in use by any Document Type, the zone cannot be deleted.

 - ungroup a cluster of child zones contained within the selected zone (child zones are used for OMR purposes).

Other Options

Skip recognition on back pages

Selecting this option will skip patch code detection on back pages, leading to increased performance during capture.

Preview

The preview function allows the user to select an image and view the Patch Code on that image.

 - Load a sample page containing barcodes to determine the types of barcodes present and check for functionality.

 - Initiate detection. Each patch code will be listed in order of appearance on the page next to their type and values.


Blank Pages

Enable Blank Page Detection

Select this option if the user desires to detect the existence of blank pages. Once Blank Page Detection is enabled, the program will attempt to determine whether the page is blank as images are captured. NOTE: If the scanner supports blank page detection at the hardware level it may provide a slight performance advantage to perform this function at that level. This option ONLY enables blank page detection. If the user wishes to delete blank pages as they are captured those options which are enabled by default are located in the Capture and Quality Assurance Configure steps of the Document Type Configuration (Step 8 of 9) workflow section of this manual.

The program has two methods to try to determine if a page is blank; both can be used interchangeably, either separately or together. If both are selected, a page will be determined to be blank if either of the 2 methods determines that the page is blank.

Blank Page Detection Settings

Use Blank Page Detection Algorithm (slower)

Using this method the program uses an advanced blank page detection algorithm to determine whether a page is blank depending on the content of the page such as type written text versus random pixels. This method is the slower of the two but may only be noticeably slower on scanners running in excess of 140 pages per minute.

Use Blank Page Detection Threshold Method

Using this method the program checks the size of what the file will be on disk and if it is less than the size specified the page is considered blank. The threshold can be set individually for both front and back of pages. This method may require some trial and error when setting document type because a number of factors can have an effect on the size of a page. Such factors may include the capture resolution (dpi), the compression method for saved documents, and whether capturing in black and white, grayscale, or color.

NOTE: Although this method can always be set for both front and back page detection, be aware when Import or Auto Import is the Capture method only (Front) is used.

NOTE: When enabling Blank Page Detection to remove blank pages from your documents, the user can choose to store the deleted Blank Pages in a separate folder for later review. This option is set in Capture Configure step of the Document Type Configuration/Workflow section of this manual.


Characters (OCR/ICR)

The recognition engines used for Optical Character Recognition and Intelligent Character Recognition throughout the workflow are configured here.

Select Default Recognition Language

Supported languages are listed for selection. NOTE: This applies only to the RecoStar engine.

Zone Settings

The recognition engine used to perform OCR/ICR functions outside of the OCR module is set to either GlyphReader or RecoStar. Different recognition engines may be desirable depending on the emphasis for accuracy versus speed.


Marks (OMR)

The recognition engine used for Optical Mark Recognition is configured here.

Mark Determination

Mark Threshold

This feature checks that the marked weight of a zone box is greater than the base weight by at least the defined amount. Move the slider to set a number between 0 – 1000.

This setting is used to counteract the effect of noise and stray marks on the scoring process. If a user does not mark any box in a zone, without a mark threshold set, the box with the most random noise will be reported as marked. If the mark threshold is set high enough that the weight from noise on the page doesn’t cross the threshold, zones that have no actual mark will be reported as blank.

Difference Threshold

This feature ensures that the reported marked weight of a zone box differs from the other boxes in the same zone by a defined minimum amount. Move the slider to set a number between 0 – 1000.

This works to improve accuracy when multiple boxes in a zone have marks present. For example: If the marks are being made with a felt tip marker and the user drags the marker across all boxes in a zone, without this setting the engine would report a mark in the box that reports the darkest mark, which very likely is a small variation from the marks in the other boxes, and may not be easily identifiable to the eye.

Ignore Base Weight values when evaluating marks

Selecting this option causes the base weight of the zone to be set to “0” and ignore the template image loaded into the document type configuration. This setting is recommended for use when drawing zones on a color image, however the scanned production documents will have the color dropped out.

Miscellaneous Options

Store image used for mark recognition as alternate page file

Stores a copy of the image used for OMR zone definition as the alternate page.

Run mark recognition against an alternate page file

If an alternate page is present, it is used as a baseline to compare the actual image against.

Alternate Page File Tag

Assign a tag to the OMR image file being saved as an alternate page file.

Mark Tags

Assign Tags to Marks

When enabled, you may use the following options:

  • By Weight
  • By Net Weight
  • By Adjusted Net Weight
  • By Scaled Weight.

Different tags can be assigned depending on conditions defined by the user.

On this page:

Related Pages: