Skip to content

Latest commit

 

History

History
428 lines (312 loc) · 26.4 KB

Release notes.md

File metadata and controls

428 lines (312 loc) · 26.4 KB

Release notes

5.0.0

BlinkInput v5 Release Announcement

As of this version, BlinkInput SDK is fully compatible with other Microblink SDKs, which means that you can use it with other Microblink SDKs in the same application.

New features:

  • We have full support for Apple Silicon!

Cocapods support change:

  • We are updating our Cocoapods with .xcframework format from now on. Please download Cocoapods 1.10.0 or newer to use our SDK.

Major API changes:

  • To ensure compatibilty with other Microblink SDKs, we have reprefixed all classes. All classes have MBI prefix instead of MB
  • We have renamed framework from Microblink to BlinkInput
  • We added error callback when setting license keys on MBIMicroblinkSDK
    • You will be getting error callback and reason why you could not unlock SDK - see MBILicenseError
  • We moved all resources inside framework, we are not shipping bundle anymore.

Minor API changes:

  • We renamed MBRecogitionMode to MBIRecognitionDebugMode in MBIRecognizerCollection.
  • Swift:
    • We renamed all sharedInstance to shared
    • All enums are now Int
    • All unsigned integers are now Int
  • We've replaced Using time-limited license! warning with Using trial license! warning. The warning message is displayed when using a trial license key.
  • We've deprecated MBDetectorRecognizer — use MBIDocumentCaptureRecognizer instead.
  • We've deprecated MBPdf417Recognizer — use MBIBarcodeRecognizer instead.

Bug fixes:

  • We've fixed a bug in IBAN parsing which occasionally read the reference number as part of the IBAN.

4.3.0

New features:

  • Introducing support for new framework format - XCFramework:
    • it contains all the necessary device and simulator architecture slices
    • no neeed to remove simulator slices before distributing your app to the App Store as described here
    • for now, only available from this repo, not available on Cocoapods
  • We improved document detection with MBDocumentCaptureRecognizer
  • Added presets for camera - Preset1080p, and 4K; Optimal always chooses the highest quality
  • We are now delivering the complete list of open source dependencies used in the SDK. Please check the open-source-software-used directory

Bug fixes:

  • Large memory consumption is due to 4K video session on all 4K eligible iPhones; we introduced new camera presets 1080p and 4K, so to reduce your app memory consumption set camera preset on 1080p or 720p

4.2.0

New features:

  • we added the option to disable Microblink logs in the console output. Use MBLogger class and conform to MBLoggerDelegate in AppDelegate. To disable logs, implement delegate method - (void)log:(MBLogLevel)level format:(const char *)format arguments:(va_list)arguments. Be careful with this option. We need full log outputs from the application for support purposes. In case of having problems with scanning certain items, undesired behavior on the specific device(s), crashes inside SDK or anything unmentioned, we will need a full log from your side. If you disable Microblink logging, you won't be able to provide us this information. Hence support might be limited.
  • added support for capturing cropped images (without data extraction) of documents of any format:
    • use MBDocumentCaptureRecognizer and MBDocumentCaptureOverlaySettings
    • use MBDocumentCaptureOverlayViewController, which is designed for taking high resolution document images and guides the user through the image capturing process. It can be used only with MBDocumentCaptureRecognizer

Improvements:

  • improved VinParser:
    • added support for Renewal Identification Number (RIN) - DMV California format
  • all 4K iPhones now use 4K resolution as default

Minor API changes:

  • methods pauseScanning and resumeScanningAndResetState in MBRecognizerRunnerViewController do not return anymore BOOL
    • use isScanningPaused to check if scanning is paused
  • Swift Module has been renamed from MicroBlink to Microblink

Bug fixes:

  • fixed Torch activation for all iOS versions
  • fixed OpenGL code which was fragile and sensitive to crashes if used from multiple threads
  • fixed issue where recognizer's result state would not be the same as recognizer's runner state after finished scanning
  • fixed issue when pressing immediately close button on overlay view controllers would freeze for couple of seconds SDK

4.1.0

  • Updates and addtitions

    • updated overlay view controllers with new icons for close and torch buttons
    • switched to using HTTPS in podspec for Cocoapods
    • enabled capturing high resolution camera frames: When custom UI integration is performed, use - (void)captureHighResImage:(MBCaptureHighResImage)highResoulutionImageCaptured on MBRecognizerRunnerViewController When using provided scan overlay view controllers, high resolution full camera frames taken at the moment of successful scan are returned if this option is enabled through MBOverlaySettings.
    • Added support for checking if scanning is unsupported for camera type on MBRecognizerRunnerViewController
    • Added reconfigureRecognizers method to MBBaseOverlayViewController, enabling it on all it's subclasses

  • Minor API changes

    • renamed MicroBlink.framework to Microblink.framework
    • renamed MicroBlink.bundle to Microblink.bundle
    • isScanningUnsupportedForCameraType: is now class method of MBMicroblinkSDK

  • Bugfixes

    • templating recognizers no longer execute callbacks with valid state once they are valid on every frame even if nothing is 'detected'
    • fixed a crash which happened when scanning region was set before overlay view controller loaded, but after it was initialized
    • fixed missing init in MBDotsResultSubview for Swift
    • fix memory issue while using current frame grabber
    • updated overlay view controllers for iPhone X Series
    • Fixed bug where SDK crashed with exception when the user wanted to use custom resource bundle
    • Calling reconfigureRecognizers before showing camera now correctly applies supplied recognizers
    • Fixed autorotation of overlays and MBRecognizerRunnerViewController
    • Fixed localization issues with some overlays
    • Fixed constraint errors on MBDocumentVerificationOverlayViewController
    • Various other bug fixes and improvements

4.0.0

  • new API, which is not backward compatible. Please check README and updated demo applications for more information, but the gist of it is:
    • PPScanningViewController has been renamed to MBRecognizerRunnerViewController and MBCoordinator to MBRecognizerRunner
    • PPBarcodeOverlayViewController has been renamed to MBBarcodeOverlayViewController
    • previously internal MBRecognizer objects are not internal anymore - instead of having opaque MBRecognizerSettings and MBRecognizerResult objects, you now have stateful MBRecognizer object that contains its MBResult within and mutates it while performing recognition. For more information, see README and updated demo applications
    • introduced MBFieldByFieldOverlayViewController that can be used for easy integration of the field-by-field scanning feature (previously known as segment scan)
    • introduced MBProcessor concept. For more information, check updated code samples, README and this blog post
  • new licence format, which is not backward compatible. Full details are given in README and in updated applications, but the gist of it is:
    • licence can now be provided with either file, byte array or base64-encoded bytes

2.3.0

  • Updates and additions:

    • Microblink.framework is now a dynamic framework. The change is introduced because of the following reasons:
      • isolation of code
      • smaller binary size
      • better interop with third party libraries (such as Asseco SEE Mobile Token)
    • improved Screen shown when Camera permission is not granted:
      • fixed crash which happened on tap anywhere on screen
      • close button can now be removed (for example, if the scanning screen is inside UINavigationController instance)
      • header is now public so you can instantiate that class if needed
    • updated PPUiSettings with new features:
      • flag showStatusBar which you can use to show or hide status bar on camera screen
      • flag showCloseButton which you can use to show or hide close button on camera screen. By default it's presented, but when inside UINavigationController it should be hidden
      • flag showTorchButton which you can use to show or hide torch button on camera screen.
    • deprecated PPHelpDisplayMode. You should replace it with a custom logic for presenting help inside the application using the SDK.
    • renamed internal extension method with namespace so that they don't interfere with third party libraries
    • added standard tap to focus overlay subview in all default OverlayViewControllers. Also added it as a public header.
    • PScanningViewController now has a simple method to turn on torch
    • simplified PPOcrLayout class (removed properties which were not used)
    • internal switch to new build system using cmake. This allows faster deployments and easier updates in the future
    • added support for hungarian parsers in segment scan
      • account number parser
      • payer ID parser
    • added support for slovenian parsers in segment scan
      • reference parser
    • all recognizer results (classes that derive PPRecognizerResult) now have annotated nullability for their getters. Some of them used to assume non-null, while still returning nil sometimes. This has now been corrected and all getters are _Nullable
    • added support for scanning IBANs that contain spaces and dashes
    • added support for scanning IBAN from Georgia in Segment Scan
    • added Belgian account number check to IBAN parser
    • added Slovak Data Matrix codes
    • added property allowResultForEveryFrame in PPScanSettings which can be used when using Direct API to force calling didOutputResults: callback for every frame
    • added feature to enable frame quality estimation when using Direct API (by exposing property estimateFrameQuality)
    • added logging of the SDK name when the license key is invalid for easier troubleshooting
    • added a property which you can use to set a custom location for resources. For example, if you would like to avoid using Microblink.bundle as resources bundle, you can set a different one in PPSettings object
    • added PPSimNumberRecognizer
    • added Generic parsing in TopUpOcrParser
    • added designated initializers to all PPOcrParserFactory objects
    • added play success sound method to PPScanningViewController protocol
    • added Barcode Recognizer PPBarcodeRecognizerResult and PPBarcodeRecognizerSettings
    • deprecated PPBarDecoderRecognizerResult and PPBarDecoderRecognizerSettings. Use Barcode Recognizer
    • deprecated PPZXingRecognizerResult and PPZXingRecognizerSettings. Use Barcode Recognizer
    • PPBlinkOcrRecognizerResult and PPBlinkOcrRecognizerSettings are now deprecated. Use PPDetectorRecognizerResult and PPDetectorRecognizerSettings for templating or PPBlinkInputRecognizerResult and PPBlinkInputRecognizerSettings for segment scan
    • added reading of mirrored QR codes
    • introduced GlareDetector which is by default used in all recognizers whose settings implement GlareDetectorOptions:
      • when glare is detected, OCR will not be performed on the affected document position to prevent errors in the extracted data
      • if the glare detector is used and obtaining of glare metadata is enabled in MetadataSettings
      • glare detector can be disabled by using detectGlare property on the recognizer settings
    • added PPQuadDetectorResultWithSize which inherits existing PPQuadDetectorResult
      • it's subclasses are PPDocumentDetectorResult and PPMrtdDetectorResult
      • returns information about physical size (height) in inches of the detected location when physical size is known
    • CFBundleShortVersionString is now updated with each release

  • Bugfixes:

    • fixed frame quality issue in PPimageMetadata. Previously it was always nan if used after image getter
    • fixed Torch button on default camera overlays. Previously it never changed state after it was turned on
    • Fixed crash when the user tapped anywhere on the view controller presented when camera permission wasn't allowed
    • fixed warning message when language is set to something other than @en, @de and @fr and @cro
    • fixed crash on start in swift if custom UI was used to handle detector results
    • fixed a problem which caused internal recognizer state not to be reset when using the scanner for the second time with the same PPCoordinator instance
    • fixed ocrLayout getter in PPBlinkOcrRecognizer which previously returned nil
    • fixed an issue which caused camera settings to be reset each time PPCoordinator's applySettings method was called. This issue manifested, for example, by automatically turning off torch after successful scan in SegmentScan
    • fixed redundant log warnings in setting language ("Trying to set language to nil, returning") and CameraManager ("Should not have been observing autofocus")
    • fixed issue with resuming camera when user is first asked for camera permission. This manifested as sometimes camera going black
    • fixed issue with blurred camera display when PPCoordinator instance was reused between consecutive scanning sessions
    • fixed crashed which happened when multiple instances of PPCoordinator were used simultaneously (one being terminated and one starting recognition). This most commonly happened when after scanning session, a new view controller was pushed to a Navigation View Controller, when the user repeated the procedure a number of times (five or more).
    • fixed issue with Direct API which disabled processing
    • fixed crash when multiple QR code-based recognizers were used together
    • fixed bug which caused didOutputResults: not to get called in DirectAPI
    • fixed case sensitivity in class & file naming
    • fixed issue which sometimes caused scanning not to be started when the user is asked for camera permission (first run of the app)
    • fixed rare crash which Camera paused label UI being updated on background thread
    • fixed incorrect handling of camera mirror when using front facing camera
    • fixed crash which sometimes happened when presenting help screens (if PPHelpDisplayModeAlways or PPHelpDisplayModeFirstRun were used)
    • fixed crash in QR code which happened periodically in all recognizers
    • fixed autorotation of overlay view controller
    • fixed scanning return result type of PPDetectorRecognizerSettings when initialized with PPMrtdDetectorSettings - returning PPMrtdDetectorResult

  • Improvements in ID scanning performance:

    • improved IBAN parser
    • improved amount parser
    • improved Croatian Reference number parsing:
      • trailing whitespace is removed from result when using Segment scanning
    • improved TopUpParser: added option to enable all prefixes at the same time (generic prefix)
    • added suport for 14 digits long sim numbers in addition to existing lengths (12, 19, 20)
    • TopUp scanning improvements
    • improved reading of pdf417 barcodes having width:height bar aspect ratio less than 2:1
    • date parsing improvements
    • added parsing of curly brackets

  • Changes in Samples:

    • added libz to all samples to prevent linker errors (caused by slimming down the SDK)
    • samples updated to use new dynamic framework
    • added a build phase in each sample which removes unused architectures from the dynamic framework
    • samples updated for XCode 9
    • all Swift samples are updated to Swift 4

2.2.0

  • iOS updates:
    • Aded Slovenian ID recognizer
    • Added parser for mobile coupons
    • Added frame quality property to PPImageMetadata
  • iOS bugfixes:
    • Fixed issue where Templating API wasn't working as expected on some devices.
    • Fixed issue with string localizations

2.1.2

  • iOS fixes:
    • CFBundleSUpportedPlatforms removed from Info.plist files
    • Applying affine transformation to PPQuadrangle now correctly assigns points.
    • When using both Direct API and PPCameraCoordinator, scanning results will now be correctly outputted to PPCoordinatorDelegate and PPScanningDelegate respectively
    • Fixed crashes related to camera permissions and added dummy view when camera permission is disabled
    • Fixed issues related to topLayoutGuide on iOS6
    • Improved performance of Date parser

2.1.1

  • iOS bugfixes:
    • Detectors are now built correctly. Incorrect builds caused Templating API not to work as desired

2.1.0

  • iOS updates:

    • Added option to mirror camera frames in 'PPCameraSettings'
    • Added VIN parser

  • iOS bugfixes

    • Fixed deadlock when 'processImage:' is called from main thread
    • Fixed setting custom PPOcrEngineOptions

2.0.0

  • iOS updates:

    • Implemented PPCameraCoordinator. PPCameraCoordinator assumes the role of PPCoordinator from previous versions while new PPCoordinator is used for Direct API (image processing without camera out management).
    • Increased speed of scanning for barcode type recognizers.
    • Implemented PPImage. When using Direct API you can wrap UIImage and CMSampleBufferRef into PPImage to ensure optimal performance.
    • Improved performance of Direct API. In addition, you can now use Direct API with your own camera management without any performance drawbacks.
    • Added method isCameraPaused to PPScanningViewController.
    • Added option to fllip input images upside down for processing with cameraFlipped property of PPCameraSettings.
    • Implemented PPViewControllerFactory for managing creation of PPScanningViewController objects.
    • PPImageMetadata now contains PPImageMetadataType property, which describes which image type was outputted.
    • Added VIN (Vehicle idendification number) parser.
    • Mirrored (both horizontally and vertically) images can now be processed by setting PPScanSettings/PPImage properties (latter if using custom camera management).

  • Implemented templating API

    • Templating API allows creation of custom document scanners, linking specific parsers to specific locations on detected documents

  • iOS bugfixes:

    • New Direct API fixed possible deadlocks when sending large amounts of data

1.2.0

  • iOS bugfixes:

    • fixed potential deadlock when multiple instances of PPCoordinator objects are instantiated

  • Implemented templating API

    • Templating API allows implementing custom document scanners, linking specific parsers to specific locations on detected documents

  • Added Regex parser

    • Regex parser allows you to create your custom ocr parser factory

  • PPOverlayViewController changed the way Overlay Subviews are added to the view hierarchy. Instead of calling addOverlaySubview: (which automatically added a view to view hierarachy), you now need to call registerOverlaySubview: (which registers subview for scanning events), and manually add subview to view hierarchy using addSubview: method. This change gives you more flexibility for adding views and managing autolayout and autoresizing masks.

  • Better Swift interoperability

    • Support for modules
    • Added nullability annotations

  • Framework is now distributed as a .framework + .bundle, instead of .embeddedframework. This helps keep resources in a separate "namespace", and avoids mistakes

  • The library inside the framework is now static library. This makes it easier to include the library inside other libraries

  • Added bitcode support for Xcode 7

1.1.1

  • Added support for using BlinkOCR as a camera capture API. To do that, implement the following

    • When initializing the PPCoordinator object, don't add any PPRecognizerSettings to scanSettings.
    • Use settings.metadataSettings.currentVideoFrame = YES to capture current camera frame
    • Implement scanningViewController:didOutputMetadata: callback to obtain PPImageMetadata objects with camera frames.

  • As a reminder - you can process video frames obtained in that way using direct API method processImage:scanningRegion:delegate:

1.1.0

  • Added support for Barcode recognition using the following recognizers:

    • PPPdf417Recognizer for scanning PDF417 barcodes
    • PPBarDecoderRecognizer for Code 39 and Code 128 1D barcodes
    • PPZXingRecognizer for all other 1D and 2D barcode types

  • Barcode scanning can be used at the same time with OCR scanning, or completely separate. Use different scanSettings.

  • Improved video frame quality detection: now only the sharpest and the most focused frames are being processed. This improves quality of the results, but at a slight expense of processing time

  • Frame quality estimation can now be enabled using PPScanSettings frameQualityEstimationMode property:

    • when set to PPFrameQualityEstimationModeOn, frame quality estimation is always enabled
    • when set to PPFrameQualityEstimationModeOff, frame quality estimation is always disabled
    • when set to PPFrameQualityEstimationModeDefault, frame quality estimation is enabled internally, if the SDK determines it makes sense

  • iOS 9 introduced new app multitasking features Split View and Slide Over. When the scanner is on screen and one of those features are used, iOS automatically pauses the Camera (this behaviour is default as of iOS 9 beta 5). This SDK version introduces new setting in PPUISettings class, called cameraPausedView, where you can define the UIView which is presented centered on screen when this happens.

  • Known issue on iOS 9: if you use Autorotate overlay feature (settings.uiSetttings.autorotateOverlay), present PPScanningViewController as a modal view controller, and support Split View iOS 9 feature, then autorotation of camera overlays isn't correct. The best way is to opt-out of Split View feature, and wait for SDK fix when iOS 9 comes out of beta.

  • PPScanningViewController methods pauseScanning, isScanningPaused, and resumeScanningAndResetState: should now be called only from Main thread, and they are effective immediately. E.g., if pauseScanning is called and there is a video frame being processed, result of processing of that frame will be discarded, if resumeScanningAndResetState: isn't called in the meantime.

  • Added support for PPCameraPresetPhoto camera preset. Use this if you need the same zoom level as in iOS Camera app. The resolution for video feed when using this preset is the same as devices screen resolution.

1.0.0

  • BlinkOCR now supports several new fonts, which especially benefit receipt scanning use case.

    • Use them by specifying a whitelist of CharKeys in PPOcrEngineOptions

    • New fonts are:

      • PP_OCR_FONT_CHAINPRINTER
      • PP_OCR_FONT_HYPERMARKET
      • PP_OCR_FONT_PRINTF
      • PP_OCR_FONT_TICKET_DE_CAISSE

  • You can now access two more properties of PPOcrChar objects

    • font - to get information about the font of the char
    • variants - to get a set of all alternative values for a specific char (note: a variant is defined as a combination of a unicode value + font, so there might be the same characters multiple times in the variants set - each time with different font)

  • PPOcrEngineOptions now have the ability to turn off line grouping (collecting adjacent chars into lines) with lineGroupingEnabled property.

  • Changes in Direct processing API

    • You are no longer required to call PPCoordinator initializeRecognizers and PPCoordinator terminateRecognizers.
    • Instead, initializeRecognizers is called lazily on first call of processImage method.
    • terminateRecognizers is called automatically in PPCoordinator destructor

  • Added support for autorotation of PPScanningViewController. To support autorotation, use PPScanningViewController's new properties autorotate and supportedOrientations

  • countless tweaks and fixes in OCR engine

0.9.6

  • Added didOutputMetadata callback method to PPOverlayViewControllers

  • Fixed bug which caused Overlay events to be called when direct (static) OCR processing is used. Direct OCR processing now only reports events to PPScanDelegate instance.

  • Fixed bug which caused image property of PPImageMetadata not take the image orientation into account

  • Added Metadata for obtaining images that are processed for OCR

0.9.5

  • Updates in OCR engine performance

  • Added method in PPOcrEngineOptions to completely disable image processing inside BlinkOCR

0.9.4

  • Added direct processing API which you can use to perform OCR on UIImage objects.

  • Added NoCamera-sample project which shows how to use direct processing API

  • You can now specify OcrEngine options to RawOcrEngineParser object. - in the options, you can customize OCR engine parameteres, like char whitelist used, minimal and maximal line sizes, etc.

  • Updated OCR models with support for more special characters like currency symbols

0.9.3

  • This version uses a new license key format. If you had a license key generated prior to v0.9.3, contact us so we can regenerate the license key for you.

  • PPCoordinator class now exposes fewer public methods and properties.

  • PPScanningViewController's methods resumeScanning and resumeScanningWithoutStateReset merged into one resumeScanningAndResetState:. - All calls to resumeScanning replace with resumeScanningAndResetState:YES. - All calls to resumeScanningWithoutStateReset replace with resumeScanningAndResetState:NO

0.9.2

  • Naming changes in API (see Transition guide)

    • PPBaseResult renamed to PPRecognizerResult
    • PPBaseResult subclasses renamed accordingly
    • PPOcrResult renamed to PPOcrLayout

  • Each PPRecognizerResult now has implemented description method for easier debugging

  • Fixed orientation handling for case when overlay autorotates.

  • Scanning region is now a property of Scanning view controller, and overlay view controller now delegates to this property.

0.9.1

  • API consolidation, Bugfixes and improvements

0.9.0

  • Initial functionality added for Raw text scanning, Price and Iban parsing.
  • Initial documentation added
  • Implemented Two sample apps
    • One for simple integration showing how OcrRecognizer can easily be included in the app
    • One with custom UI for FormScanning with Price, IBAN and Reference number parsing