Document Reader SDK 5.2 — August 4, 2020
New result types implemented to simplify SDK integration:
combined text results from all sources in
RPRM_ResultType_Text combined graphics results from all sources in
RPRM_ResultType_Images aggregated processing and validation statuses in
RPRM_ResultType_Status. Added capability to read both sides of bank cards in
“CreditCard” scenario (multipage support). Added correct depersonalization support for all platforms with applying blur over the field areas. Depersonalization configuration options extended with
allTextFields and allGraphicFields. Document type recognition
Improved with new NN under the hood. Smaller size, better quality results.
New functionality implemented to better differentiation between document series
based on document data.
Added support for Arabic, Bengali, Korean, Hindi, Sinhala, Tamil, and Thai languages.
Refactored to optimize and remove redundant code and make it smaller in size.
Fixed issue with Azeri Ə character returned Cyrillic instead of Latin.
Added new parsers:
Saudi Arabia visa.
Kuwait ID card,
Romania ID card.
Improved reading on images with color noise over MRZ.
Fixed issue with setting validity for the fields that were not actually validated.
Fixed issue with Qatar ID card code reading.
Text data parsing and validation
Added content validation for Romanian CNP number (personal number).
Added date validation for
ft_DateOfBirthOfWife & ft_DateOfBirthOfHusband fields. Added separate calculated
ft_Category field when processing bank cards number presenting issuer payment system. Fixed issues with parsing registration stamp data into specific fields for Russian Federation domestic passports.
Fixed issue with visual vs MRZ comparison for values that contain ‘.
Fixed issue with
^ symbol in ft_Surname, ft_GivenNames and ft_Surname_And_Given_Names in lexical analysis results. Image QA
Image quality assessment NNs (focus and glares) were updated for better result quality, smaller size, and improved performance.
Added capability to generate integral image, that consists of many images taken in series, and does not contain glare. Works only if
“doublePageSpread” option is off. Glare presence over text fields prevents process series finish in OCR enabled scenarios.
Added support for patterns check in Axial light (new HKG Id card).
Fixed issue with high intensity luminescence objects in pattern schecks.
Fixed issue with performing checks that were not selected.
Added new eVisualFieldType enum members:
ft_CVV Added new
eRPRM_ResultType enum members:
RPRM_ResultType_Images New structures added:
TRfidOrigin Restored filling of text and graphic field names in result structures.
Code quality improved with lots of small issues fixed.
Core RFID SDK
Fixed JSON serialization for
RFID_ResultType_RFID_OriginalGraphics result type. Hardware support
Fixed issue with document detection in video detection mode on 70x4M-5A devices.
Fixed issue with image calibration compensation for 70x9-10 device added in 5.1.
Added capability to return all images captured by device if
“processParam”.”noImageFilters” provided in parameters via RPRM_Command_Set_ProcessParametersJson command call. Fixed issue with loading and storing calibration data on Linux.
Fixed issue with
RPRM_ResultType_Portrait_Comparison not available in __ResultTypeAvailable method. Fixed issue with serial port RTS and CTS toggling on device activity for IATA VCOM integration.
Added the ability to select a mode of the recognition process (
Auto), start video mode using the Capture button (
CaptureVideo), manual mode for processing a single frame (
CaptureFrame). Added the ability to set the maximum time for document recognition, after which the recognition process ends automatically (
timeout). In some scenarios, you can manually correct the borders of the detected document when using the automatic recognition mode via the
“showScanner” function ( manualCrop). In some scenarios, you can manually correct the borders of the detected document when using the
“recognizeImage” function ( manualCrop). Added the ability to limit the document tilt angle (
perspectiveAngle). Added the ability to set the minimum allowed DPI of a document in the image (
minDPI). Added the ability to display metadata while reading a document (
showMetadata). Added the ability to get an integral image (
integralImage). Added SDK localization in 8 new languages: Italian, Polish, Arabic, Vietnamese, Czech, Portuguese, Thai, Turkish.
Fixed minor issues.
Added a new function
getTextFieldValueByTypeAndSource to quickly get values from the results by specifying eVisualFieldType and eRPRM_ResultType. Added a new function
getTextFieldValueByTypeAndSource to quickly get values from the results by specifying eVisualFieldType, eRPRM_ResultType values and original properties. Added a new function
getTextFieldValueByTypeAndSource to quickly get values from the results by specifying eVisualFieldType and original properties. In
DocumentReaderResults, added a new property holding time spent on reading the document RFID chip ( elapsedTimeRFID). In
DocumentReaderResults, added a new property holding time spent on document recognition ( elapsedTime). In
DocumentReaderResults, added a new property "processingFinishedStatus" indicating the current status of the recognition process ( ProcessingFinishedStatus). In
RGLDocumentReaderValue, added a new property indicating recognition probability (0% -100%) ( probability). Added the ability to change the quality of the video obtained from the device camera during document recognition (
cameraSize). Redesigned the UI for reading the document RFID chip.
Redesigned the error structure when working with the SDK, now most errors are returned as "Throwable".
Fixed image serialization issue for
DocumentReaderGraphicField. Fixed the instructional animation when using the
"CreditCard" scenario. Fixed memory leak caused by incorrect work of
Added a new function
getTextFieldValueByType to quickly get values from the results by specifying RGLFieldType and RGLResultType. Added a new function
getTextFieldValueByType to quickly get values from the results by specifying RGLFieldtype, RGLResultType and original properties. Added a new function
getTextFieldValueByType to quickly get values from the results by specifying RGLFieldtype and original properties. In
RGLDocumentReaderResults, added the morePagesAvailable status indicating that the document contains other pages. In
RGLDocumentReaderResults, added the time spent on reading the document RFID chip ( elapsedTimeRFID). In
RGLDocumentReaderResults, added the time spent on document recognition ( elapsedTime). In
RGLDocumentReaderResults, added a new property "processingFinishedStatus" indicating the current status of the recognition process ( RGLProcessingFinishedStatus). In
DocumentReaderValue, added a new property indicating recognition probability (0%-100%) ( probability). Added new values to the
RGLFieldTypeFt_CVV Redesigned the error structure when working with the SDK, now most errors are returned as
"NSError" objects. Web Service API
OpenAPI specification published on
GitHub with examples. npm package available. Java client published on
GitHub with examples. maven package available. Python client published on
GitHub with examples. pipy package available. Implemented complete backward compatibility with old API (
/GetTransactionResult & /GetInfo missing endpoints added).
/ping endpoint results extended to include license information. Document Reader Application
Added capability to select active RFID device by its name in “RFID chip” tab in “Options” dialog.
Added capability to select active camera device by its name instead of its index in “Live comparison” tab in “Options” dialog.
Fixed issue with creation of scanned list, where data was rewritten by previous scan in video detection mode if quick MRZ results mode enabled.
Fixed issue with hiding application icon from taskbar if “Hide minimized” options is selected, and main window is closed (when “Allow window closing” is disabled).
Fixed issue with detecting light type when loading images from folder, now filenames that contain
“ir” or “uv” will not be considered unless exact match with full filename. Documentation updated.
ActiveRFIDDeviceIdx to set or return the index of the currently active connected RFID device (integer, read/write). Added property
AvailableRFIDDevices to return the RFID device name specified by index in the Index parameter from the list of available devices. Added property
AvailableRFIDDevicesCount to return the number of available devices in the system (integer, read only). Added property
LiveCameraCount to return the number of available web cameras in the system (integer, read only). Added property
LiveCameraDevices to return the web camera name specified by index in the Index parameter from the list of available cameras. Added property
LiveCameraTimeout to set or return the capture timeout in milliseconds for obtaining a live portrait from the selected web camera during the live comparison process. 5000 by default (integer, read/write). Fixed issue with
CustomProcess method not returning correct output. Enums updated.
Windows Desktop API
Fixed issue with COM registration for x64 package version.
New drivers package version 1.8.11 with Regula camera driver version 10.5.9.8 included.
ICAO PKD master list version 152 included in RFID SDK component.
Fixed missing Latin fields, written in italic, in all Bulgarian documents.
246 countries and territories / 8808 documents included.
130 new documents added.
For details see
Supported documents list.
Back to top