Configurations for Barcode Recognition

Prev Next

If there are a large number of imported documents, it can make sense to specify the barcodes to be recognized in a separate template. This template can be assigned to a document processing template.

Best practice for barcode types

The selection of a suitable barcode type ensures high performance and precision and depends on the processed documents. In a barcode template you can set which barcodes should be recognized and read.

  • For a high recognition rate, it is advisable to select as few barcode types as possible, ideally only one.

  • 1D Barcodes: Code 39 and Code 128 are considered the most stable linear barcodes in terms of damage and distortion

  • 2D Barcodes: Data Matrix is considered the most stable 2D barcode in terms of damage and distortion, and also the most compact (before Aztec, QR and PDF417).

  • If your typical document contains many tables, use Aztec and not DataMatrix. If the document contains many circular objects, use DataMatrix and not Aztec. Type PDF417 should be avoided if possible.

  • Recognition quality is often better when using grayscale images. In the case of black and white images, you should ensure that the resolution is high enough to get a sharp image of the barcode. Always ensure high contrast. If the black lines or squares contain many white or light pixels, this reduces the quality of the recognition.

Barcodes: Advanced Settings

For some documents it may be necessary to define further properties for Barcode Recognition in the barcode template, namely under Barcode > More options. This is not necessary for most documents and is not recommended. If you do need to define other properties, however, please contact DocuWare Support, who will be happy to help you set up the right properties for your specific documents.

Below is a list of the most important properties and possible values:

Property

Possible values

Automatic Mode

  • true, false

  • Default value: false

This mode is useful if you are not satisfied with the general recognition quality. If you set this option to true, the recognition quality is improved, but the barcode recognition is also slowed down considerably. Depending on the image content, the process can take up to 30 seconds per page, although the average is around 3 seconds per page, ten times longer than normal. The ExpectedBarcodes property must then also be carefully adjusted, as the algorithm may not search for other barcodes once the specified number is reached.

BarcodeTypes

  • Any combination of barcode types

  • Minimum selection: one entry

  • Default value: all types

Limit the list to the types that are actually used. This setting is most effective and will improve performance and precision, so it should be adjusted first.

BoundingBoxRatio

  • Default value: 0.02

  • This ratio is calculated by dividing the shorter by the longer side of the Bounding Box. This is to filter out results that have a very low ratio (barcode is very "thin") because they are not valid barcodes.

Encoding

  • If a barcode is read incorrectly and incorrect characters are displayed, you can activate a character encoding that matches the barcode for each code page supported by your system.

  • Examples: Arabic signs: 1256; Kyrillic signs: 1251; Japanese Signs: 932; if no other code page works: 65001.

Erode

  • true, false

  • Default value: false

If images with low contrast are used, this setting improves the quality of the recognition by artificially increasing the contrast. Do not use this option if the image has sufficient contrast, especially not for scans with a low DPI value, as this would worsen the result further.

FilterDataFlags

  • true, false

  • Default value: false

Removes all data flags (e.g. information about sub-symbology and coding) and returns the value that should be coded according to ISO/IEC 8859-1. This setting can be useful when using a special sub-symbology that cannot be configured and if the pure value meets your expectations exactly.

ISO threshold

  • -1.00, 0.00 to 4.00

  • Default value: -1.00

ISO 15415 is used to analyze the 2D barcodes DataMatrix, QR, MicroQR, Aztec.
ISO 15416 is used to analyze these 1D barcodes: EAN13, EAN8, UPCA, UPCE, Interleaved2of5, Code11, Code39, Code93, Codabar, Code128, MSI, Pharmacode, Telepen, Standard2of5, RSS
In both cases, the test returns a value between 0.00 and 4.00. If the returned value is below the configured threshold, the barcode is discarded. 0.00 means that the barcode is likely to be rejected, while 4.00 means that a barcode that is easy to analyze was found. If many reject barcodes are found, you can increase the value slightly, but it should not exceed 0.10. The value -1.00 means that the test is deactivated.

RenderResolution

  • 200, 300

  • Default value: 300

This is only important if you use PDFs that contain text and are therefore seen as native. If easily recognizable and large barcodes are used, this value can be reduced to 200 to improve performance.

UseChecksum

  • true, false

  • Default value: false

The following barcode symbologies have an optional checksum: Code 39, Industry Standard 2 of 5, Interleaved 2 of 5.
If UseChecksum is set on real barcodes without a checksum or with an incorrect checksum, these are discarded. If a barcode with a checksum is read while the option is set to false, the checksum appears as an additional character at the end of the barcode unit.

Experimental Settings

ConfidenceThreshold

  • 0.00 - 1.00

  • Default value: 0.95

ExpectedNoBarcodes

  • Any whole number equal to or greater than zero

  • Default value: 0

It is advisable to consult DocuWare Support before making any changes. You should set this value to at least the number of barcodes that appear on average on one page, but not significantly higher. The recognition process tries to find the configured number of barcodes. If the number is too small, those that are very easy to recognize will still be found. Barcodes that are more difficult to recognize could however be overlooked once the expected number is reached. It is not recommended to set this value to 99 as this may additionally impair performance.

QualityThreshold

  • 0.00 - 1.00

  • Default value: 0.20

It is advisable to consult DocuWare Support before making any changes.

TradeOff

  • BestSpeed, Balanced, BestQuality

  • Default value: BestSpeed

If barcodes are not recognized (especially smaller barcodes), switch to BestQuality. This will decrease performance and increase the average number of reject barcodes recognized. Changing the TradeOff can improve the result, but can also worsen it, so caution is advised with this setting. If you change the TradeOff, the TradeOff value should be adjusted (the higher the TradeOff, the lower the quality threshold).