Difference between revisions of "OCR templates"
m (→Add field zones) |
m (→Add field zones) |
||
| Line 30: | Line 30: | ||
|- | |- | ||
|'''Type''' | |'''Type''' | ||
| − | |The field type ( String, Date etc | + | |The field type ( String, Date, etc. ) |
|- | |- | ||
|'''Property''' | |'''Property''' | ||
| Line 36: | Line 36: | ||
|- | |- | ||
|'''Pattern''' | |'''Pattern''' | ||
| − | |Java pattern ( take a look at http://docs.oracle.com/javase/7/docs/api/java/util/regex/Pattern.html ) | + | |Java regular expression pattern ( take a look at http://docs.oracle.com/javase/7/docs/api/java/util/regex/Pattern.html ) |
|- | |- | ||
|'''Custom''' | |'''Custom''' | ||
| − | | | + | |Commands for processing images before OCR. It can contain several commands, one per line ( example: '''/usr/bin/convert ${fileIn} -matte ( +clone -fuzz 50% -transparent red ) -compose DstOut -composite -negate ${fileOut}''' ) |
|- | |- | ||
|'''OCR''' | |'''OCR''' | ||
| − | | | + | |OCR command line ( example: '''/usr/bin/tesseract ${fileIn} ${fileOut} digits''' ) |
|- | |- | ||
|'''Use to recognise''' | |'''Use to recognise''' | ||
| Line 48: | Line 48: | ||
|- | |- | ||
|'''Zone''' | |'''Zone''' | ||
| − | | | + | |Here, the cropped image is shown. |
|} | |} | ||
Revision as of 10:29, 8 August 2014
 |
Since OpenKM 6.4.2 the plug-in system of OpenKM allows you to expand quickly the functionality offered by the platform, extending the available OCR field parsers without having to rebuild the system to add/change the existing functionality. Refer to Extend OCR field parsers if you need to extend OCR field parsers feature on 6.4.2. |
OCR Templates allows to create zonal OCR templates which allows to recognise and extract estructured text from scanned images.
|
Images should be scanned at least at 200 dpi of resolution to get good text recognition from OCR engine. |
Template creation
Open OCR template administration option.
Then click on the 
new ocr template icon
Fill the form and click create button.
Add field zones
| Field | Description |
| Name | Template field name |
| Type | The field type ( String, Date, etc. ) |
| Property | The property group parameter name ( example okp:property.name ) |
| Pattern | Java regular expression pattern ( take a look at http://docs.oracle.com/javase/7/docs/api/java/util/regex/Pattern.html ) |
| Custom | Commands for processing images before OCR. It can contain several commands, one per line ( example: /usr/bin/convert ${fileIn} -matte ( +clone -fuzz 50% -transparent red ) -compose DstOut -composite -negate ${fileOut} ) |
| OCR | OCR command line ( example: /usr/bin/tesseract ${fileIn} ${fileOut} digits ) |
| Use to recognise | Indicate if a field is also used to identify the document or only for extraction purpose. |
| Zone | Here, the cropped image is shown. |
Then click on the 
fields icon
Fill the form and select the zone and click create button.
Test OCR template
Click on the 
check icon and will be shown all the data fields extracted by zone.
Recognise testing
From main OCR template list, click 
recognise icon.
Fill the form selecting some scanned image to test recognision.
Click recognise button.
Enable Zonal OCR data capture
OCR data capture can be enabled from profiles and automation.
Profiles
To enable zonal OCR data capture from profiles should be enabled the check OCR data capture.
Automation
Automation is divided in two operations, validations and actions.
There's a validation called IsOCRDataCaptureFile which validates if OCR data capture engine supports the image format.
There're two actions OCRDataCapture and AddOCRDataCaptureToWizard. OCRDataCapture capture data and store to metadata. AddOCRDataCaptureToWizard enables end user wizard to see live ocr data capture process.
For more information take a look at Automation.
