Skip to content

Release 3.2 — June 9, 2022

Mobile API

iOS

Added

  • Added RFSLivenessProcessingContentView and RFSLivenessRetryContentView for user interface customization puproses.
  • Added configuration properties to LivenessConfiguration and FaceCaptureConfiguration: torchButtonEnabled, closeButtonEnabled.

Changed

  • Renamed LivenessConfiguration and FaceCaptureConfiguration property from cameraSwitchEnabled to cameraSwitchButtonEnabled.

Fixed

  • Fixed MatchFaces crash where MatchFacesRequest was given an empty UIImage object. The request does not support nil images, however, it was possible to pass UIImage() with empty image data.
  • Fixed multithreading issues with video recorder. Rare app crashes when closing liveness are fixed now.
  • Fixed crashes in the RFSCopyCVPixelBuffer method. In some cases row by rows copying was used.

Android

Added

  • Added the registerProcessingFragment method to LivenessConfiguration to set custom processing fragment.
  • Added the setCloseButtonEnabled and setTorchButtonEnabled methods to Configuration.
  • Added the setForceToUseHuaweiVision method to MatchFaceConfiguration and LivenessConfiguration to force using Huawei Vision library.

Web API

Core Functionality

Detection

  • Added attributes: shades, blur, brightness, noise, eyes status (closed/open).

Matching

  • Changed the recognition model version from 3.1 to 4.0.

Liveness 3D

  • Updated the face quality models.
  • Updated the occlusion model version.
  • Updated the model for assessing that an object is three-dimensional.
  • Updated the liveness verification checklist.
  • Added the default and login modes.

February 1, 2022

iOS

Added

  • Added new instruction texts on the retry screen for when Liveness fails.
  • A new API endpoint is used for RFSMatchFacesRequest that introduces new data and structures available in the response, such as results and detection. See RFSMatchFacesResponse.
  • Added a new class RFSMatchFacesImage that represents an input image for RFSMatchFacesRequest. This class is used instead of RFSImage to provide new attribution for features supported by the new API endpoint.
  • RFSMatchFacesComparedFace now has new properties: image, faceIndex, and face.
  • RFSMatchFacesRequest now has a new property thumbnails. It makes responses contain a cropped thumbnail image of a detected face.
  • Added a new error code RFSMatchFacesErrorImagesCountLimitExceeded for RFSMatchFacesError.
  • Added a new class RFSMatchFacesSimilarityThresholdSplit for managing matched and unmatched RFSMatchFacesComparedFacesPair results by given the similarityThreshold value.
  • Added a new class RFSMatchFacesDetection that represents detection results on an input image.
  • Added a new class RFSMatchFacesDetectionFace that represents face detection information.
  • Added a new class RFSPoint that represents a two number value (X, Y).
  • Added a new image type RFSImageTypeExternal that represents an image from an unknown source.
  • Added a new liveness configuration option isLocationTrackingEnabled that defines whether the liveness request sends a location of a device. The location is used only when permissions are granted to the host application and the location is available.
  • Added localization support for languages:
    • Croatian (hr)
    • Danish (da)
    • Hungarian (hu)
    • Indonesian (id)
    • Ukrainian (uk)
    • Romanian (ro)
    • Slovak (sk)
    • Norwegian (nb)

Fixed

  • Fixed a bug where a current attempt was not incremented properly and caused LivenessConfiguration.attemptsCount to have no effect in some cases where Vision would fail to validate a face on the first frame.
  • Fixed a bug where notification hint Fit your face in the oval was interfering with other statuses causing the animation to jump back and forth between the correct text and the former.

Deprecated

  • RFSMatchFacesRequest deprecated the similarityThreshold property. See RFSDeprecated.h.
  • RFSMatchFacesResponse deprecated the matchedFaces, unmatchedFaces properties. See RFSDeprecated.h.
  • RFSMatchFacesComparedFace deprecated the imageType, position properties. See RFSDeprecated.h.
  • RFSImage deprecated the tag property. See RFSDeprecated.h.

Changed

  • RFSMatchFacesRequest now uses the RFSMatchFacesImage type as an input image.
  • Liveness processing logic has changed. Now the detection oval height changes instead of the camera's zoom.
  • Renamed the enum RFSLivenessError to RFSLivenessErrorCode.
  • Renamed the enum RFSMatchFacesError to RFSMatchFacesErrorCode.
  • Renamed the enum RFSFaceCaptureError to RFSFaceCaptureErrorCode.
  • Renamed the class RFSComparedFace to RFSMatchFacesComparedFace.
  • Renamed the class RFSComparedFacesPair to RFSMatchFacesComparedFacesPair.
  • Renamed the enum case RFSImageTypeLiveWithDoc to RFSImageTypeDocumentWithLive.

Removed

  • Removed error codes RFSMatchFacesErrorNotInitialized, RFSMatchFacesErrorCommandNotSupported, RFSMatchFacesErrorCommandParamsReadError from RFSMatchFacesError as they were not informative for the public API and may be handled as a general RFSMatchFacesErrorProcessingFailed instead.
  • Removed RFSComparedFacesPairBackendError and RFSComparedFacesPairError. RFSMatchFacesError is used instead.
  • Removed the tag property from the RFSMatchFacesComparedFace class.
  • Removed camera force / auto focus for Liveness when the screen is touched.

Android

Added

  • Added new instruction texts on the retry screen for when Liveness fails.
  • A new API endpoint is used for MatchFacesRequest that introduces new data and structures available in the response, such as getResults and getDetections. See MatchFacesResponse.
  • Added a new class MatchFacesImage that represents an input image for MatchFacesRequest. This class is used instead of Image to provide new attribution for features supported by the new API endpoint.
  • MatchFacesComparedFace now has new methods: getMatchesFaceImage, getFaceIndex, and getDetectionFace.
  • MatchFacesRequest now has a new method setThumbnails that makes responses contain a cropped thumbnail image of a detected face.
  • Added a new error code IMAGES_COUNT_LIMIT_EXCEEDED for MatchFacesError.
  • Added a new class MatchFacesSimilarityThresholdSplit for managing matched and unmatched MatchFacesComparedFacesPair results by given the similarityThreshold value.
  • Added a new class MatchFacesDetection that represents detection results on an input image.
  • Added a new class MatchFacesDetectionFace that represents face detection information.
  • Added a new liveness configuration option isLocationTrackingEnabled that defines whether the liveness request sends a location of a device. The location is used only when permissions are granted to the host application and the location is available.
  • Added localization support for languages:
    • Croatian (hr)
    • Danish (da)
    • Hungarian (hu)
    • Indonesian (id)
    • Ukrainian (uk)
    • Romanian (ro)
    • Slovak (sk)
    • Norwegian (nb)
  • Added support for Huawei services via the Huawei Visio library.

Changed

  • Renamed FaceCaptureResultCodes to FaceCaptureErrorCode.
  • Renamed enum values in ImageType.
  • com.regula.facesdk.model.results.MatchFacesResponse moved to com.regula.facesdk.model.results.matchfaces.MatchFacesResponse.
  • com.regula.facesdk.model.results.FaceCaptureResultCodes moved to com.regula.facesdk.model.results.FaceCaptureErrorCode.
  • com.regula.facesdk.model.results.FaceCaptureResultCodes moved to com.regula.facesdk.model.results.FaceCaptureErrorCode.
  • com.regula.facesdk.model.results.MatchFacesErrorCodes moved to com.regula.facesdk.model.results.MatchFacesErrorCode.
  • Changed the FaceCaptureException return value from the FaceCaptureErrorCode enum in the getErrorCode method.
  • Changed the LivenessErrorException return value from the LivenessErrorCode enum in the getErrorCode method.
  • Changed the MatchFacesException return value from the MatchFacesErrorCode enum in the getErrorCode method.
  • Changed the Image return value from the ImageType enum in the getImageType method.
  • Changed the LivenessResponse return value from the LivenessStatus enum in the getLiveness method.
  • MatchFacesRequest now uses the MatchFacesImage type as an input image.
  • Liveness processing logic has changed. Now the detection oval height changes instead of the camera's zoom.
  • Renamed the class com.regula.facesdk.model.results.ComparedFace to com.regula.facesdk.model.results.matchfaces.MatchFacesComparedFace.
  • Renamed the class com.regula.facesdk.model.results.ComparedFacesPair to com.regula.facesdk.model.results.matchfaces.MatchFacesComparedFacesPair.
  • Renamed the case IMAGE_TYPE_LIVE_WITH_DOC to DOCUMENT_WITH_LIVE in the ImageType enum.

Removed

  • Removed the tag property from the MatchFacesComparedFace class.
  • Removed error codes NOT_INITIALIZED, COMMAND_IS_NOT_SUPPORTED, COMMAND_PARAMS_READ_ERROR from MatchFacesErrorCode as they were not informative for the public API and may be handled as a general PROCESSING_FAILED instead.