Release notes

Version 9.14.1 (2024-10-18)

  • Adds Portrait.fromBuffer and Portrait.toBuffer methods.

  • Adds FaceTracker.DetectFaces method.

Version 9.14.0 (2024-10-16)

New features

  • New passive presentation attack detection model.

  • New user instructions in PortraitProcessor module.

  • Add compression artifact remover.

API changes

  • Reviewed API of the PortraitProcessor module

  • Added PortraitInstruction enumeration

  • Modified PortraitStatus enumeration

  • Removed FacePad.ColorPadMinimumIod

  • Added Portrait.padScore

  • Added Portrait.lookStraightScore

  • Renamed PortraitQualityAttributes to PortraitQualityCheckpoints.

Bug fixes

  • Bad interpolation of the mask in FaceAnalyser.SegmentFace

Version 9.13.2 (2024-08-09)

  • [Python] Added Python 3.10 package on Windows and Linux

Version 9.13.1 (2024-07-24)

  • [Windows] fixed cudart.dll dependency (which fixes [Python] ImportError: Missing DLL)

Version 9.13.0 (2024-07-22)

  • [Python] Reworked Image import/export from Numpy array :

    • Removed support from numpy array in fromBuffer functions

    • Added specific function Image.from_numpy()

    • Added support for various PixelFormat and fix crashes with unsuitable data

    • Fixed bug in to_numpy() function with some pixels formats.

  • Fixed several bugs in FaceAnalyser.segmentFace, segmentBackground and applyMask

  • Fixed irrelevant backend logs

  • Internal updates : codegen 3.37.0, Image 3.10.0

  • [Java] Added support for macOS.

  • [Python] Added support for macOS.

Version 9.12.0 (2024-06-04)

  • Added smart card license support.

  • Minor fixed in Python wrapper.

Version 9.11.0 (2024-05-15)

API changes

  • Renamed FaceSegmenter1A and function SegmentFace to Face BackgroundSegmenter1A and SegmentBackground

  • Added detailed options in ProcessingUnit enum for specific backends.

  • Automatic model selection (among those loaded) in all modules, removing the need to call SetProcessingUnit or SetModel.

New features

  • New model FaceSegmenter2A and function SegmentFace in FaceAnalyser for complete face parts segmentation. All available parts are in new enum FaceSegmentationLabel.

  • New builds available for different new GPU backends:

    • Windows/Linux : default (OpenCL) & CUDA

    • Android : default (CPU Only) & GPU (OpenCL, Vulkan)

    • Apple : default (Metal)

  • All models now available for GPU use. Warning : This is experimental, some models might be unstable on some backends.

Bug fixes

  • Face detection default threshold set to 50 (best for models 4X).

  • Python wrapper on linux : fix CUDA lib mandatory dependency

Version 9.10.0 (2023-11-10)

New features

  • Added PortraitProcessor.removeBackground method.

  • Added support for Grayscale8Bits destination pixel format in Image.fromRawBuffer().

  • Image.toRawBuffer now accepts Bgr24Bits pixel format as ouput.

API changes

  • Merged PortraitProcessor.analysePortrait and PortraitProcessor.updatePortrait methods to PortraitProcessor.createPortrait.

  • Renamed processing options in the PortraitProcessor module.

Bug fixes

  • Fixed issue in Image.downscaleTo pixel format update.

  • Fixed issue in Image.getPixelDepth when picture is empty.

  • Fixed issue in Image.copyTo when picture is empty.

Version 9.9.2 (2023-11-02)

  • Rename Portrait.emotion property to Portrait.expression.

  • Improved documentation of the FaceAttributes structure.

Version 9.9.1 (2023-10-18)

  • Changed default PortraitProcessor.mouthOpenThreshold value from 5 to 35.

  • Fixed PortraitProcessor.detectRedEyes method.

  • Adds documentation and sample code for the PortraitProcessor module.

  • Skips color PAD in PortraitProcessor if model is not loaded.

Version 9.9.0 (2023-10-13)

New features

  • Improved presentation attack detection in PortraitProcessor module.

API changes

  • PadStatus enumeration added.

  • Portrait.isPresentationAttack replaced by Portrait.padStatus.

  • PortraitQualityAttributes member names changed for clarity.

Version 9.8.0 (2023-09-27)

New features

  • New unified multi-language documentation.

  • New model for face PAD: FaceAttackSupportDetector3A.

  • Inference engine updated.

API changes

  • FaceAnalyser.applySegmentationMask method renamed FaceAnalyser.applyMask and now takes 3 parameters to set the background color.

  • FaceAnalyser.applyAlphaMask method added => returns a BGRA image with transparent background.

  • FaceAnalyser.threadCount property added.

  • License module renamed FaceLicense.

  • FaceLibrary.checkLicense methods moved to FaceLicense module.

  • FaceLibrary.getVersionEx method added.

  • FaceTemplateBufferType enumeration removed.

  • Some enumerations renamed in Dart (ex: FaceTemplateFormat) and Java (ex: PixelFormat).

  • New Swift API.

  • Updated Python package API :

    • Camel case for methods and attributes, upper case for enumerations.

    • Renamed package to lower case.

Version 9.7.0 (2023-08-09)

  • New wrapper for Python

  • Added build for NXP iMX (Edge Edition)

  • PortraitProcessor module:

    • Improved portrait quality computation - Bug fixes

  • FaceAnalyser module:

    • Fix under-exposure and over-exposure default sensibility values

Version 9.6.1 (2023-06-14)

  • Fixed bug in PhotographicAttributes computation for overexposure and underexposure scores

Version 9.6.0 (2023-05-30)

  • Added new background removal feature (segmentFace & applySegmentationMask) in FaceAnalyser

Version 9.5.1 (2023-04-21)

  • Fixed issue in ColorBased Pad

Version 9.5.0 (2023-03-21)

  • Added Face Detector 4A

  • Added new Portrait Processor module used to facilitate the creation of portraits in various scenarios such as enrolment of face verification. The module performs the following tasks:

    • Face detection and tracking with prediction models

    • Landmark detection

    • Head pose estimation

    • Subject’s position verification

    • Template encoding and updating

    • Additional optional computations (age estimation, background uniformity computation, emotion detection, eye gaze detection, eye opening detection, face mask detection, gender estimation, glasses detection, hat detection, make-up detection, mouth opening estimation, occlusion detection, photographic attributes computation, presentation attack detection, red-eye detection, smile detection)

    • Verification of compliance on many quality attributes

    • Unified quality score computation

  • Fixed left/right eye in occlusions

  • Added method to update a FaceTemplate

  • Added FaceAnalyzer.computePoseWithLandmarks() method for more accurate pose estimation.

  • Added DetectedFace.getExpandedBounds() method.

Version 9.4.3 (2023-03-28)

  • Fixed memory automatic dispose in internal Python wrapper

Version 9.4.2 (2023-03-10)

  • Fixed issue in ColorBased Pad

Version 9.4.1 (2023-03-03)

  • Fixed issue with FaceTracker using model FaceDetector4B

Version 9.4.0 (2023-02-08)

  • Added Face Detector 4B

    • better accuracy than Face Detector 3B with roughly the same speed

  • Fixed issue in Face Attack Support Detector

  • Fixed issue in Tracker IDs uniqueness

  • Fixed issue in DetectedFace getDistanceToCamera() method

  • Fixed issues in FaceImage (…)To() methods

  • Reduced binary size for dynamic librairies

  • Added on Image :

    • Replace Image.Convert() by Image.toRawBuffer().

    • Added support for YV12 pixel format.

    • Added Image.Compare().

Version 9.3.4 (2022-12-21)

  • Correct Face Analyser code documentation

Version 9.3.3 (2022-11-24)

  • Update internal Image dependency.

    • Corrected JPEG2000 support

Version 9.3.2 (2022-11-24)

  • Fixed issues in Image rotateTo(), resizeTo() and flipTo() api

Version 9.3.1 (2022-10-17)

  • Fixed multiple issues in Swift wrapper

Version 9.3.0 (2022-09-12)

  • Added FaceTracker module

    • Realtime multiple faces tracking

    • Internally using face encoder to ensure consistency of tracked identities

  • Added Java finalizers on SDK objects

  • Added Dart finalizers on SDK objects

    • id3Face SDK now requires Dart >= 2.17

  • Corrected invalid return from FaceLibrary.getVersion()

Version 9.2.1 (2022-06-17)

  • Added minimum IOD parameter in Face Encoder

    • Allow to bypass or reduce the minimum IOD check of 30px for face encoding introduced in 9.2.0 (not recommended to do so)

Version 9.2.0 (2022-06-15)

  • Major changes in API:

    • Added face indexer

    • Updated occlusion detection model to 2A

    • Removed deprecated Moiré algorithm due to poor performance

    • Added ICAO check methods: ComputeNoise(), CheckColorizedImage()

    • Updated over-exposure and under-exposure measure algorithm

    • Replaced brightness field with OverExposure and UnderExposure in PhotographicAttributes

    • Replaced FaceContrast with FaceDynamicRange in PhotographicAttributes

  • Updated inference library:

    • Fixed memory leaks in model unloading

  • Added minimum IOD for color-based PAD

  • Added minimum IOD of 30 pixels for face encoder

  • Fixed memory leaks in function DetectFaces()

  • Deprecated models:

    • Deprecated occlusion detection model 1A

    • Deprecated Moiré detection model 1A

Version 9.1.5 (2022-05-17)

  • Added Ubuntu 18.04 x64 GPU build

Version 9.1.4 (2022-05-16)

  • Fixed invalid list size in DetectedFaceList_GetIdList() api

  • Fixed issues when setting thread count in FaceDetector and FaceEncoder

Version 9.1.3 (2022-05-09)

  • Fixed desktop edition license check

Version 9.1.2 (2022-05-04)

  • Added Ubuntu 18.04 x64 build

Version 9.1.1 (2022-04-26)

  • Fixed issue in eye openness and eye redness algorithms

  • Fixed issue in background analyser for some image sizes

  • Fixed issue when using USB based licenses on windows

Version 9.1.0 (2022-04-12)

  • Major changes in API:

    • Added color based face PAD algorithm 2A

    • Added face attack support detector algorithm 2A

    • Added face detector 3C

    • FaceTemplate and DetectedFace _Save() APIs renamed to _ToFile()

    • Added Warmup() APIs for FaceDetector and FaceEncoder

    • Added GetVersion() API

  • Updated inference library:

    • From 10 to 40% speedup (depending of model, input size, …)

    • Reduced RAM usage for very large inputs

  • Added javadoc jar

  • Fixed invalid values for eye gaze estimation

    • Eye gaze angle values now returned as degrees

  • Fixed invalid values for background uniformity estimator

  • Fixed potential issue in FaceMatcher when invalid templates were provided

  • Fixed execstack warning when using the Face SDK in a Java Linux environment

  • Deprecated models:

    • Deprecated color based face PAD algorithm 1A

    • Deprecated face attack support detector algorithm 1A

Version 9.0.4 (2022-03-14)

  • Fixed issue when using several objects of the same kind in parallel

  • Fixed potential crash in eye gaze estimation

Version 9.0.3 (2022-03-02)

  • Fixed periocular encoder inference

  • Fixed issue on dart wrapper when running on iOS devices

Version 9.0.2 (2022-02-28)

  • Fixed search candidate list order

  • Fixed face encoding quality estimator (replaced 2A with 3A)

Version 9.0.1 (2022-02-17)

  • Fixed potential crash when calling some objects from different threads

Version 9.0.0 (2022-02-09)

  • Major changes in API:

    • Moved encoding quality estimator to face encoder module

    • Modified parameter getters/setters for each module

    • Added face template buffer type

    • Added face eye gaze structure

    • Added face background uniformity structure

    • Removed segmenter API

    • Merged all libraries in single binary for all targets

  • Added face detector 3A and 3B

    • Major improvement in detection accuracy

  • Added face encoder 9A and 9B

    • Major improvement in recognition accuracy

  • Added dart wrapper

  • Added swift wrapper

  • Updated image library:

    • Minor API changes

  • Updated inference library:

    • Increased inference speed on all target platforms

    • Added TensorRT support

    • Cuda Toolkit (Linux and Windows) dependencies updated to:

      • Cuda 11.5

      • Cudnn 8.3.1

  • Updated license library:

  • Optimized activation methods

  • Standardized API

  • Add all wrappers

Version 8.9.5 (2021-08-06)

  • Modified license check APIs for android: does not need activity pointer anymore

  • Secured calls to detected face functions

  • Fixed potential memory leaks when calling face template from buffer several times in a row

Version 8.9.4 (2021-07-28)

  • Added expression sensibility parameter to FaceAnalyser

  • Corrected issue in template BIT export

Version 8.9.3 (2021-06-15)

  • Fixed face depth PAD in API

Version 8.9.2 (2021-06-02)

  • Updated id3Image dependency to 1.7.1:

    • Added BGRA 32 bits pixel format support

Version 8.9.1 (2021-05-04)

  • Fixed issue on ComputeEyesAttributes Java API

Version 8.9.0 (2021-04-02)

  • Added iOS and MacOS support (Preview version) :

    • Distributed as a combined universal XCFramework with C API (can be directly called from ObjectiveC code)

    • iOS :

      • Architectures armv7, armv7s and arm64 + simulator

    • MacOS

      • Architectures x86_64 and arm64

  • API changes:

    • Added BIT and BDT template export api

  • Misc :

    • Modified _ToBuffer() APIs return codes:

      • now returns “Insufficent Buffer” error instead of “Success” when provided too small export buffers

Version 8.8.0 (2021-03-04)

  • New models:

    • Added RGB-based face PAD algorithm version 1A

    • Added RGB-based face PAD blurriness detection algorithm version 1A

    • Added RGB-based face PAD Moiré detection algorithm version 1A

    • Added RGB-based face attack support detection algorithm version 1A

    • Added face segmenter algorithm version 1A

  • API changes:

    • Added uniformity background algorithm in FaceAnalyser

  • Misc:

    • Fixed sharpness function

    • Updated id3 Inference dependency to 2.5.3:

      • Protected parallel calls for inference on GPU

    • Updated id3 Image dependency to 1.7.0:

      • Added ImageBuffer API

      • Fixed correctGamma API

Version 8.7.4 (2021-03-03)

  • Correct android samples issues

Version 8.7.3 (2021-02-16)

  • Updated id3 Inference dependency to 2.5.2:

    • Corrected potential issue when performing face tracking in a loop

Version 8.7.2 (2021-02-09)

  • Updated id3 License dependency to 7.2.0:

    • Changed hardware ID on Jetson Nano

      • New method of computation, requires to regenerate licence files

      • Jetson Nano hardware IDs now end with “09”

Version 8.7.1 (2021-02-04)

  • API changes:

    • Added FaceLicense_Reactivate APIs in C, C# and Java

Version 8.7.0 (2021-01-29)

  • API changes:

    • Added FaceTemplateDict object API

    • Added API to create a DetectedFace object from values

    • Added API to get portrait landmarks from a DetectedFace object

    • Added parameters to the API to get portrait bounds from a DetectedFace object

    • Replaced FaceTemplateList by FaceTemplateDict in the FaceMatcher search API

  • Misc:

    • Updated id3 Image dependency to 1.4.0:

      • Added error codes

      • Fixed C# fromBuffer API

    • Updated id3 License dependency to 7.1.0:

      • Modified id3License_GetLicenseType() behaviour

      • Added id3LicenseType enum

    • Return nullptr handles when initializations fail

Version 8.6.3 (2021-01-15)

  • Fixed license activation in C# when using console inputs

Version 8.6.2 (2021-01-14)

  • Fixed license activation in C#

Version 8.6.1 (2021-01-07)

  • Fixed id3 Inference dependency to 2.4.3:

    • Fixed GPU support for some platforms

  • Fixed photographic attributes crash

Version 8.6.0 (2020-12-11)

  • New models:

    • Added age estimation algorithm 1A

    • Added face depth-based presentation attack detection (PAD) algorithm version 2A

      • Increases the accuracy of the face liveness detection for depth images

    • Added face expression classifier algorithm version 1A

      • Allows to detect if a person has a neutral expression or not for an ID photograph

    • Added face occlusion detection algorithm 1A

      • Allows to detect if the eyes are hidden behind tinted glasses, or if the mouth and nose are covered by a mask

  • Bug fixes:

    • Fixed eye attributes classification algorithm

      • The outputs of the algorithm have changed, be careful when updating to adapt the thresholds to the new documentation

    • Fixed photographic attributes potential crash

    • Fixed issue with FingerTemplate_ToBufferCC() with encoder 8A

    • Updated id3 Inference dependency to 2.4.2

      • Corrected potential crashes introduced in 8.5.0 for some models

  • Deprecated models:

    • Depreciated face depth-base PAD algorithm version 1A

Version 8.5.0 (2020-11-25)

  • Supported targets:

    • Updated id3 Inference dependency to 2.4.1

      • Added Windows, Linux and Jetson Nano GPU CUDA support for models FaceEncoder8A and FacePeriocularEncoder1A

  • Bug fixes:

    • Fixed empty search list to return success instead of invalid template format

    • Fixed detected face list copy function

  • Misc:

    • Updated id3 Image dependency to 1.3.1

      • Allowed buffers to be too large for toBuffer methods

    • Updated id3 License dependency to 7.0.3

      • Activation CLI tool now handles inputs by environment variables

Version 8.4.1 (2020-10-19)

  • Updated id3 Inference dependency to 2.2.2

    • Corrected inference time regression on windows

  • Updated id3 Image dependency to 1.2.3

  • Fixed rescale API for DetectedFaceList

Version 8.4.0 (2020-10-09)

  • New models:

    • Added periocular algorithm version 1A

      • It is recommended to use this encoder when users wear chirugical masks

    • Added face quality estimation algorithm version 2A

      • The face quality score is now better correlated with the matching performances

    • Added face attribute estimation algorithm 2A

      • Lighter model and more precise attributes

  • Misc:

    • Modified API for the ComputeAttributes function in the FaceAnalyser module

  • Bug fixes:

    • Corrected error handling issues in template I/O functions

    • Removed noise in console outputs when using the SDK

  • Deprecated models:

    • Depreciated face quality estimation algorithm version 1A

    • Depreciated face attribute estimation algorithm 1A

Version 8.3.2 (2020-09-23)

  • Fixed id3 Image dependency to 1.2.1

Version 8.3.1 (2020-09-21)

  • Added rescaling API for a detected face

  • Fixed detector bug:

    • there is no internal resizing anymore

    • min and max head sizes are deprecated

    • it is up to the app developer to handle a resizing if necessary for real-time speed (see samples)

  • Updated id3 Image dependency to 1.2.1

Version 8.3.0 (2020-09-16)

  • Added Raspberry Pi 4 support (Raspberry Pi OS armv7l)

  • New models :

    • Added face encoder 8B for 1-to-1 embedded applications

      • Encoder8B is released under specific licensing and can only be used with the associated license package

    • Updated face landmarks estimation model to FaceLandmarksEstimator_2A (lighter and more precise)

    • Updated head pose estimation model FacePoseEstimator_1A (lighter, more precise and now only requires detection results)

  • New APIs :

    • Added API to retrieve Compact Card templates

    • Added APIs to retrieve id3 license files from id3 server (C / C# / Java)

  • Misc

    • Added dependency to id3 Image Library 1.1.0

    • Updated license dependency to id3 License 7.0.1

      • Support for “online” type licenses

      • New linux and jetson nano hardware codes - will require to re-generate licenses on these platforms

      • Added CLI activation tool on all platforms

        • Retrieve license file using id3 credentials, activation key or serial key

        • Print local hardware code

        • Display a license file information

      • CheckLicense() functions now require a path/buffer to a license file

    • Reduced binaries size for all libraries

  • Bug fixes :

    • Corrected issue when using Face Detector in a multithread context

    • Corrected issues in C# samples post build events

Version 8.2.1 (2020-08-24)

  • Fixed android delivery

Version 8.2.0 (2020-08-03)

  • Updated Jetson Nano build to Nvidia Jetpack 4.4

  • Added C# capture sample

  • Bug fixes :

    • Corrected issue in detector module multithread behaviour

    • Corrected private visibility for EyesAttributes java fields

    • Corrected noise in android logcat when using detector or encoder_v8a

Version 8.1.2 (2020-07-24)

  • Bug fix :

    • Corrected issue in android template extraction for models 7s/7a on android

Version 8.1.1 (2020-07-17)

  • Bug fixes :

    • Fixed issue in id3Inference.so linux x64 dependencies

    • Fixed regression in computation times on android arm64-v8a for some models

      • face_encoder_v7a/s

      • face_quality_estimator_v1a

Version 8.1.0 (2020-07-16)

  • Added Windows x86 library

  • New APIs :

    • Load model from a buffer

    • Rotate and Translate point

    • DetectedFace copy

  • Updated samples :

    • Added C# sample to create a compressed thumbnail of a face image

    • Added android capture sample

  • Bug fixes :

    • Fixed issue in image copy function

    • Fixed issue in android samples delivery

Version 8.0.0 (2020-07-01)

  • Major changes in the API. See documentation for more details.

  • Improved face detection algorithm.

  • Improved face recognition algorithm: added face encoder 8A.

  • Improved ICAO criteria classification algorithm.

  • Added face mask detection.

  • Added Jetson Nano libraries.

  • Removed FaceCapture API (will be released as a separate package).

Version 7.5.10 (2020-03-01)

  • Corrected android readme and sample on license activation issues

  • Removed depreciated android activation app apk

Version 7.5.9 (2020-02-11)

  • Change the ‘Time to live’ (TTL) value for ONVIF Discover.

  • Added functions to log Camera IP Discover.

Version 7.5.8 (2020-01-24)

  • ANDROID: Change FaceImage_nSetPixelsFromAndroidPlane() and Java FaceImage.setPixelsFromAndroidPlane() API to add support for the 3 YUV planes buffer.

Version 7.5.7 (2019-10-30)

  • Fixed crash if model file ‘ENCODER_07.BIN’ is not present even if it not used.

  • Added id3Image_ClonePicture() to creates a copy of a specific picture type of a FaceImage object.

Version 7.5.6 (2019-10-22)

  • Fixed bug with Logitech C270 Webcam.

Version 7.5.5 (2019-10-03)

  • Added java function FaceNative.setPath and getPath.

Version 7.5.4 (2019-08-08)

  • Fixed crash in id3FacePad module.

Version 7.5.3 (2019-08-01)

  • Changed FaceCaptureFragment.FaceCaptureListener.onFaceProcess() callback to onFaceProcessed().

Version 7.5.2 (2019-07-31)

  • Improve capture/detection process with Android.

Version 7.5.1 (2019-07-18)

  • Fixed bug: The Java event onFaceDetected doesn’t work.

  • Added Java function FaceTemplateDatabase.clear()

Version 7.5.0 (2019-07-16)

  • Fixed bug: The Face Processed event triggered a Face Removed event.

  • Added FaceCapture.GetFirstMulticamSlotForDevice(), FaceCapture.GetAllMulticamSlotForDevice() and FaceImage.getAllFaceBounds().

  • Addef FaceCaptureDevice.startPreview(), stopPreview() and capturePhoto() in Java.

  • Added FaceCapture java API for FaceCameraMediaType and FaceMediaFileMode.

  • Fixed FaceCapture java API bugs.

Version 7.4.0 (2019-06-18)

  • Added id3FaceEncoder_LoadModels(), id3FaceEncoder_UnloadModels() and id3FaceEncoder_SetFaceExtractionProcessingUnit().

  • Remove enum FaceEncoderStatus, replace by FaceError.

  • Changed API for id3FaceEncoder_Initialize(): remove ‘model_path’ and add ‘ExtractionProcessingUnit_Cpu’.

  • Changed FaceEncoder API: models must be loaded (and unloaded) by the user before (and after) using objects.

  • Integrated FaceCapture on Linux.

Version 7.3.5 (2019-06-26)

  • Added FaceCapture java modules.

Version 7.3.4 (2019-06-21)

  • Fixed bug in FaceImage_nGetShapes (java in Android)

  • Fixed bug in Android faces tracking

Version 7.3.3 (2019-06-20)

  • Added FaceCapture Linux and FaceCapture java modules.

  • Fixed bug in FaceImage_nToBuffer() (java)

Version 7.3.2 (2019-05-14)

  • Fixed bugs.

  • Fixed RotateAndCrop for ICAO compliance.

Version 7.3.1 (2019-04-19)

  • Fixed bugs.

  • Suppress x64_gpu folder.

Version 7.3.0 (2019-04-13)

  • Added modules FaceTemplateRecord and FaceTemplateDatabase.

  • Made FaceMatcher thread-safe.

  • Updated RealSense SDK version to 2.20.0.771.

Version 7.2.0 (2019-04-09)

  • Deleted x64_GPU folder.

  • Added functions to set extraction processing unit (CPU or GPU).

  • Changed enum FaceExtractionProcess to FaceExtractionProcessingUnit.

Version 7.1.1 (2019-03-20)

  • Fixed crash with RealSense when depth process and rotation is enabled both.

Version 7.1.0 (2019-03-12)

  • Added GPU/CPU support with the sames entry points in id3Face.dll.

  • Added getter to access quality from template.

  • The number of multicam slot is no longer limited.

  • FaceCapture license now controls the number of streams (webcam, files, ip camera) used in parallel.

  • Added FaceRecognitionProcess.

  • Added depth-based pad detection algorithm.

Version 7.0.1 (2019-02-13)

  • Fixed bug in search process. Scores were often equal

  • Fixed java fromValue function on all enum

  • Change return error code for id3FaceMatcher_AddReference() when the template format is not good then the error code is ID3_E_INVALID_FORMAT rather than ID3_E_INVALID_PARAMETER

  • Change id3FaceTemplate_GetFormat() parameters, return an id3FaceTemplateFormat enumeration.

  • Fixed bugs in id3FaceMatcher_AddReference and id3FaceTemplate_FromBuffer when the template has a wrong format

Version 7.0.0 (2019-02-05)

  • Added algo version 3.0.0 => 2 possible extraction(optim for Speed or optim for accuracy)

Version 6.10.3 (2019-02-01)

  • Fixed FaceAssembly.Initialize() when using x64_gpu subfolder in bin folder.

Version 6.10.2 (2019-01-31)

  • Fixed FaceAssembly.Initialize() when used in a multi AppDomain context.

Version 6.10.1 (2019-01-29)

  • Added Realsense autoconfiguration with the first available video format (640x480x6 fps)

  • Added support for left and right infrared stream (RealSense)

Version 6.10.0 (2019-01-28)

  • Corrected bug in Intel RealSense module.

  • Corrected lock of webcam at FaceCapture initalizing.

  • Added support for capturing the infrared stream of the camera.

  • Added infrared projector control support for the camera.

  • Added support for saving a depth picture and a infrared picture.

  • It is now possible to change value of the property CaptureBuffers between 8 to 64.

  • Added support to use infrared stream in the detect and track process.

Version 6.9.0 (2018-12-20)

  • Added android online license activation functions.

  • Corrected lacking dependency dll vcomp140.dll on windows since 6.8.0.

  • Added support for Intel RealSense depth camera.

  • Added API id3Image_SetRequestedPictureType and id3Image_GetRequestedPictureType.

  • Added API FaceDrawing.DrawMosaik

  • Added support for média player loop mode (id3FaceCameraProperty.MediaLoopMode)

Version 6.8.1 (2018-11-22)

  • Fixed detection module initialization bug with a camera.

  • Fixed crash in the streaming module when setting rotation or flip.

Version 6.8.0 (2018-11-20)

  • Added multithread mode for face extraction (~ x3.5 speedup).

Version 6.7.5 (2018-11-14)

  • Corrected bug introduced in 6.7.4 in face detection on linux and android systems.

Version 6.7.4 (2018-10-31)

  • Added FaceDrawingInfo.SetMessage()

  • Added FaceDrawingProcess.GetFaceDrawinginfoForUserID()

  • Fixed bug in quality score computation

  • Fixed bug in GetAndLockPicture(): the function now return the correct picture index

  • FaceCapture:

    • Breaking API changes :

      • Deleted FaceCapture.StopPreview(), use FaceCapture.StopAllCameras() instead

      • Deleted FaceCapture.StartPreview(), use FaceCamera.StartPreview() instead

      • Deleted FaceCapture.TakePicture(), use FaceCamera.TakePicture() instead

      • Deleted property FaceCapture.SelectedCamera, use FaceCapture.SelectCamera() instead

    • Added id3FaceCapture_SetFaceProcessedEvent (c++) and FaceProcessed (C#): event called when the tracker finish it’s process

  • C# API:

    • Added property FaceMatcher.ReferencesCount

    • Fixed flicker with the background in FaceDrawingProcess when the control is resized

  • Android API:

    • Added FaceCaptureFragment.getCameraType() function

Version 6.7.3 (2018-10-25)

  • Corrected samples

  • Fixed the crash when using the GPU

Version 6.7.2 (2018-10-24)

  • Modified FaceEncoderStatus (C++,C#,JAVA) (Success or error codes)

Version 6.7.1 (2018-10-23)

  • Added ‘data’ and ‘windows\java’ folder in samples

  • Corrected bug when using software gamma if rotation is 180°

Version 6.7.0 (2018-10-19)

  • Added quality assessment API (Warning: ComputeQuality becomes ComputeQualityMetrics)

  • Android :

    • Corrected issue when calling setCameraType() on a non initialized FaceCaptureFragment

    • Added onFaceCaptureViewCreated() interface callback

    • Corrected horizontal swap on capture

Version 6.6.0 (2018-10-15)

  • Corrected bugs in multicam process

  • Android :

    • Added support of back camera

    • Added setCameraType() function to FaceCaptureFragment to perform camera switches during capture

Version 6.5.0 (2018-10-12)

  • Optimized extract and detect on windows x64

  • Renamed QUALITY.bin model file as ICAO.bin

  • Fixed bug in rotateAndCrop

Version 6.4.1 (2018-10-08)

  • Corrected 100% CPU usage in stream capture module

  • Corrected bug with IP Camera (password and name)

  • Corrected bug with Webcam with RGB24 or I420 video format

Version 6.4.0 (2018-10-02)

  • Added Motion Detection in FaceCapture

  • Fixed stream capture issues

  • Fixed problems when started/stopped webcams

  • Added functions in id3FaceLicense API to mimic id3License API

  • Fixed bug in search on list JAVA implementation

Version 6.3.0 (2018-09-27)

  • Added API to search on list of IDs

  • Corrected bug in getEyesPosition C# API

Version 6.2.9 (2018-09-24)

  • Added getAndroidHardwareID android function.

  • Added id3License_ShowHardwareID.apk tool in android activation

Version 6.2.8 (2018-09-18)

  • Fixed missing files in the SDK package

Version 6.2.7 (2018-08-10)

  • Fixed bug when FaceID is read before the tracking thread is finished.

Version 6.2.6 (2018-08-08)

  • Corrected start/stop preview crash with IP Camera and stream.

Version 6.2.5 (2018-08-06)

  • Corrected bug in RotateAndCrop.

Version 6.2.4 (2018-07-30)

  • Added inpedendant thread for streaming capture (STREAM and IP CAM)

Version 6.2.2 (2018-07-26)

  • Changed internal storage method for template.

  • Corrected bug in FaceImage.FaceCount et FaceIDList (C/C#) when a face as been detected by not yet tracking.

Version 6.2.1 (2018-07-25)

  • Corrected bug in FaceImage.RotateAndCrop

Version 6.2.0 (2018-07-23)

  • Android changes :

    • Added support for evaluation licenses

    • Added android license manager apk

    • Added android usage samples

  • Java and Android implementations :

    • Corrected FaceImage.getHeadOrientation bug

    • Corrected FaceSpoof.getResult bug

  • FaceImage : Suppress FaceSlot and replace it by FaceID.

    • FaceID automaticaly increment if a new face is detected.

    • All API that used faceSlot now use FaceID.

    • It is possible to change the size of the list of FaceID using property MaxNumberOfFacesSlots (FaceCamera and FaceEncoder).

  • FaceDrawing new APIs :

    • FaceDrawing.DrawBitmap() to automaticaly draw all face drawing informations (use of a dictionnary is optional).

    • FaceDrawing.DrawFaceID() to draw the FaceID inside the cropping bound.

    • FaceDrawingProcess module to draw face information in a background worker. This module include the FaceID dictionnary for DrawBitmap.

  • New face spoofing detection algorithm

  • Corrected multi-thread issue with FacesData.

Version 6.1.3 (2018-07-19)

  • External sur id3.Licences V5.7

Version 6.1.2 (2018-07-05)

  • Corrected bug (lost face) in TakePicture.

Version 6.1.1 (2018-07-02)

  • Corrected FaceDrawing crash when the application is hidden in the task bar.

Version 6.1.0 (2018-06-29)

  • Added support for Matrix Vision cameras.

  • Added event for file analysis: FaceCapture.CaptureEvent with StreamAnalysePicture type.

  • Added capture support for Android.

  • Corrected bug in FaceImage for BMP image with 4 channels.

Version 6.0.4 (2018-06-28)

  • Added Capture Buffer Reset (SetProperty with automatic flag)

Version 6.0.3 (2018-06-26)

  • Added FaceImage.ColorProcess() to correct brightness, contrast and gamma.

  • Deleted support for the FaceCameraProperty_FrameRate property in webcam and IDS device.

Version 6.0.2 (2018-06-26)

  • Added support for YUYV and NV12 format for webcams.

Version 6.0.1 (2018-06-22)

  • Added software gamme correction for capture device (FaceCameraProperty_SoftwareGamma).

  • Property ‘FaceCameraProperty_Gamma’ renamed ‘FaceCameraProperty_HardwareGamma’

  • Accelerate StretchHistogram() function.

  • Removed ‘ID3_E_NO_DEVICE’ error in StartPreview().

  • Corrected drawing bug in FaceDrawing.DrawFaceImage() if color depth of the image is not 24 bits.

Version 6.0.0 (2018-06-15)

  • Great improvement in face recognition accuracy (new templates are are not compatible with previous templates created from older SDK)

  • Only one type of template available. Size is now 264 bytes.

Version 5.4.2 (2018-05-23)

  • Added support for Histogram stretching (level min and max)

  • Extended support for FrameRate (all value from min to max)

Version 5.4.1 (2018-05-18)

  • Corrected FaceCapture.SelectCamera default init for callEvent (true)

  • Moved ClearDetectedFace internal call from StopPreview to StartPreview

Version 5.4.0 (2018-05-16)

  • Added support for FileAnalysis

  • Added property descriptor for FaceDetectionEnabled, FaceDetection_MinHeadSize, FaceDetection_ConfidenceThreshold, FaceDetection_Speed and FaceDetection_Mode.

  • Added FaceCapture.CallSelectCameraEvent and FaceCapture.SelectCamera callEvent argument.

Version 5.3.0 (2018-05-14)

  • Added support for FrameRate property

Version 5.3.0.7 BETA (2018-04-16)

  • Added preload the first picture for video file

  • Added flush of decoder buffer when seeking video file

Version 5.3.0.6 BETA (2018-04-09)

  • Added multi-cam support (beta)

  • Use native FaceImage as captured pictures

Version 5.3.0.5 BETA (2018-04-06)

  • Corrected capture process

  • Corrected drawing process

  • Changed preview and capture event

  • Corrected samples

Version 5.3.0.2 BETA (2018-03-14)

  • Correct bug error 155 with IDS Imaging cameras

  • Correct playing file with 29.97 FPS

  • Suppress API id3FaceCapture_SelectDevice()

Version 5.3.0.0 BETA (2018-03-12)

  • Correct c# sample

  • Update support of event GetUserPassword for IP cameras

  • [JAVA] nGetBitrate becomes nGetColorDepth

Version 5.2.4 (2018-03-07)

  • Correct crash of sample ‘capture’

Version 5.2.3 (2018-02-27)

  • Correct crash in id3FaceEncoder_ComputeQuality

Version 5.2.2 (2018-02-21)

  • Correct bug in IDS TakePicture

  • Correct bug in FaceImage property default value for FaceCapturePreview control

Version 5.2.1 (2018-01-15)

  • Change use of camera handle under API

  • Correct bug under image format for webcam

  • id3FaceCapture_SelectDevice is deprecated, use id3FaceCapture_SelectDeviceByHandle

Version 5.2.0 (2018-01-08)

  • Add plug & play support

  • Add API function ‘id3Image_GetDetectionScore’

  • Add Linux

Version 5.1.1 (2017-12-13)

  • Canon DSLR: Force ISO Auto in preview mode

  • Canon DSLR: Fix a problem with external flash triggering

Version 5.1.0 (2017-11-14)

  • Add NVIDIA GPU support

  • Add tracking functions in C#

  • Add Java wrapper

Version 5.0.1 (2017-09-15)

  • Improvement of multi face tracking process

  • Improve tight cropped face detection

Version 5.0.0 (2017-08-10)

  • Improvement of face detection process

  • New face antispoofing module

Version 4.0.0 (2017-04-26)

  • Great improvement of face extraction and matching process

Version 3.3.3 (2017-02-23)

  • Solves a face detection issue that occured in some conditions when the subject moved while taking the picture with Canon DSLR.

Version 3.3.2 (2017-01-31)

  • Manage JPEG-2000 compression ratio

  • Bug fix

Version 3.3.0 (2016-12-05)

  • Add support for Canon DLSRs

  • Add support for IDS uEye cameras

Version 3.2.0 (2016-08-12)

  • Improvement in face quality analysis

  • New API to handle multiple face in C# and C/C++

  • Improvement in face tracking for high yaw angle

Version 3.1.5899 (2016-02-25)

  • Improvement in facial verification and identification

  • Improvement in non frontal face detection

  • Memory leaks fixes

Version 3.0.5794 (2015-11-24)

  • Improvement in facial recognition

  • Multi face tracking ( C# and C/C++)

  • Bug fixes

Version 2.2.5542 (2015-03-06)

  • Some API changes

  • Adds Java API

  • Bug fixes

Version 2.2.5030 (2014-09-29)

  • Integration of facial recognition.

Version 1.0.5030 (2013-10-09)

  • Initial release.