BlinkInput SDK for real time text recognition

BlinkInput SDK is a state-of-the-art OCR module for mobile devices. It’s OCR technology is optimized specifically for mobile devices and architectures. This allows faster results and lower error rate than regular desktop-based OCR software. BlinkInput features:

• integrated camera management
• integrated text parsing feature for fields like IBANs, prices, email addresses, urls, and many more!
• layered API, allowing everything from simple integration to complex UX customizations.
• lightweight and no internet connection required
• enteprise-level security standards
• support for barcode scanning

BlinkInput is a part of family of SDKs developed by MicroBlink for optical text recognition, barcode scanning, ID document scanning and many others.

BlinkInput powers PhotoMath app where it’s used to recognize mathematic expressions in real time.

Table of contents

• Requirements
• Quick Start
• Advanced BlinkInput integration instructions
  • Built-in overlay view controllers and overlay subviews
    • Using MBIBarcodeOverlayViewController
    • Using MBIFieldByFieldOverlayViewController
    • Using MBIDocumentCaptureOverlayViewController
    • Custom overlay view controller
  • Direct processing API
    • Using Direct API for NSString recognition (parsing)
• MBIRecognizer and available recognizers
• List of available recognizers
  • Frame Grabber Recognizer
  • Success Frame Grabber Recognizer
  • PDF417 recognizer
  • Barcode recognizer
  • BlinkInput recognizer
  • Detector recognizer
  • Document Capture recognizer
• MBIProcessor and MBIParser
  • The MBIProcessor concept
    • Image Return Processor
    • Parser Group Processor
  • The MBIParser concept
    • Amount Parser
    • Date Parser
    • Email Parser
    • IBAN Parser
    • License Plates Parser
    • Raw Parser
    • Regex Parser
    • TopUp Parser
    • VIN (Vehicle Identification Number) Parser
• Scanning generic documents with Templating API
  • Defining how document should be detected
  • Defining how fields of interest should be extracted
    • The MBIProcessorGroup component
    • List of available dewarp policies
    • The MBITemplatingClass component
    • Implementing the MBITemplatingClassifier
• The MBIDetector concept
  • List of available detectors
    • Document Detector
    • MRTD Detector
• Localization
• Troubleshooting
  • Integration problems
  • SDK problems
    • Licencing problems
    • Other problems
  • Frequently asked questions and known problems
• Size Report
• Additional info

Requirements

SDK package contains BlinkInput framework and one or more sample apps which demonstrate framework integration. The framework can be deployed in iOS 9.0 or later.

SDK performs significantly better when the images obtained from the camera are focused. Because of that, the SDK can have lower performance on iPad 2 and iPod Touch 4th gen devices, which don’t have camera with autofocus.

Quick Start

Getting started with BlinkInput SDK

This Quick Start guide will get you up and performing OCR scanning as quickly as possible. All steps described in this guide are required for the integration.

This guide sets up basic Raw OCR parsing and price parsing at the same time. It closely follows the BlinkOCR-sample app. We highly recommend you try to run the sample app. The sample app should compile and run on your device, and in the iOS Simulator.

The source code of the sample app can be used as the reference during the integration.

1. Initial integration steps

Using CocoaPods

brew install git-lfs
git lfs install

• Be sure to restart your console after installing Git LFS

• Note: if you already did try adding SDK using cocoapods and it’s not working, first install the git-lfs and then clear you cocoapods cache. This should be sufficient to force cocoapods to clone BlinkInput SDK, if it still doesn’t work, try deinitializing your pods and installing them again.

• Project dependencies to be managed by CocoaPods are specified in a file called Podfile. Create this file in the same directory as your Xcode project (.xcodeproj) file.

• If you don’t have podfile initialized run the following in your project directory.

  pod init
  

• Copy and paste the following lines into the TextEdit window:

platform :ios, '9.0'
target 'Your-App-Name' do
    pod 'PPBlinkOCR', '~> 5.0.0'
end

• Install the dependencies in your project:

$ pod install

• From now on, be sure to always open the generated Xcode workspace (.xcworkspace) instead of the project file when building your project:

open <YourProjectName>.xcworkspace

Integration without CocoaPods

-Download latest release (Download .zip or .tar.gz file starting with BlinkInput. DO NOT download Source Code as GitHub does not fully support Git LFS)

OR

Clone this git repository:

brew install git-lfs
git lfs install

• Be sure to restart your console after installing Git LFS

• To clone, run the following shell command:

git clone git@github.com:BlinkInput/blinkinput-ios.git

• Copy BlinkInput.xcframework to your project folder.

• In your Xcode project, open the Project navigator. Drag the BlinkInput.xcframework file to your project, ideally in the Frameworks group, together with other frameworks you’re using. When asked, choose “Create groups”, instead of the “Create folder references” option.

• Include the additional frameworks and libraries into your project in the “Linked frameworks and libraries” section of your target settings.

  • libc++.tbd
  • libiconv.tbd
  • libz.tbd

2. Referencing header file

In files in which you want to use scanning functionality place import directive.

Swift

import BlinkInput

Objective-C

#import <BlinkInput/BlinkInput.h>

3. Initiating the scanning process

To initiate the scanning process, first decide where in your app you want to add scanning functionality. Usually, users of the scanning library have a button which, when tapped, starts the scanning process. Initialization code is then placed in touch handler for that button. Here we’re listing the initialization code as it looks in a touch handler method.

Also, for initialization purposes, the ViewController which initiates the scan have private properties for MBIRawParser, MBIParserGroupProcessor and MBIBlinkInputRecognizer, so we know how to obtain result.

Swift

class ViewController: UIViewController, MBIDocumentOverlayViewControllerDelegate  {

    var rawParser: MBIRawParser?
    var parserGroupProcessor: MBIParserGroupProcessor?
    var blinkInputRecognizer: MBIBlinkInputRecognizer?

    override func viewDidLoad() {
        super.viewDidLoad()
    }

    @IBAction func didTapScan(_ sender: AnyObject) {

        let settings = MBIDocumentOverlaySettings()
        rawParser = MBIRawParser()
        parserGroupProcessor = MBIParserGroupProcessor(parsers: [rawParser!])
        blinkInputRecognizer = MBIBlinkInputRecognizer(processors: [parserGroupProcessor!])

        let recognizerList = [self.blinkInputRecognizer!]
        let recognizerCollection = MBIRecognizerCollection(recognizers: recognizerList)

        /** Create your overlay view controller */
        let documentOverlayViewController = MBIDocumentOverlayViewController(settings: settings, recognizerCollection: recognizerCollection, delegate: self)

        /** Create recognizer view controller with wanted overlay view controller */
        let recognizerRunnerViewController: UIViewController = MBIViewControllerFactory.recognizerRunnerViewController(withOverlayViewController: documentOverlayViewController)

        /** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
        present(recognizerRunnerViewController!, animated: true, completion: nil)
    }
}

Objective-C

@interface ViewController () <MBIDocumentOverlayViewControllerDelegate>

@property (nonatomic, strong) MBIRawParser *rawParser;
@property (nonatomic, strong) MBIParserGroupProcessor *parserGroupProcessor;
@property (nonatomic, strong) MBIBlinkInputRecognizer *blinkInputRecognizer;

@end

@implementation ViewController

- (void)viewDidLoad {
    [super viewDidLoad];
}


- (IBAction)didTapScan:(id)sender {

    MBIDocumentOverlaySettings* settings = [[MBIDocumentOverlaySettings alloc] init];

    self.rawParser = [[MBIRawParser alloc] init];
    self.parserGroupProcessor = [[MBIParserGroupProcessor alloc] initWithParsers:@[self.rawParser]];
    self.blinkInputRecognizer = [[MBIBlinkInputRecognizer alloc] initWithProcessors:@[self.parserGroupProcessor]];

    /** Create recognizer collection */
    MBIRecognizerCollection *recognizerCollection = [[MBIRecognizerCollection alloc] initWithRecognizers:@[self.blinkInputRecognizer]];

    MBIDocumentOverlayViewController *overlayVC = [[MBIDocumentOverlayViewController alloc] initWithSettings:settings recognizerCollection:recognizerCollection delegate:self];
    UIViewController<MBIRecognizerRunnerViewController>* recognizerRunnerViewController = [MBIViewControllerFactory recognizerRunnerViewControllerWithOverlayViewController:overlayVC];

    /** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
    [self presentViewController:recognizerRunnerViewController animated:YES completion:nil];

}

@end

4. License key

A valid license key is required to initalize scanning. You can generate a free trial license key, after you register, at Microblink developer dashboard.

You can include the license key in your app by passing a string or a file with license key. Note that you need to set the license key before intializing scanning. Ideally in AppDelegate or viewDidLoad before initializing any recognizers.

License key as string

You can pass the license key as a string, the following way:

Swift

MBIMicroblinkSDK.shared().setLicenseKey("LICENSE-KEY")

Objective-C

[[MBIMicroblinkSDK sharedInstance] setLicenseKey:@"LICENSE-KEY"];

License key as file

Or you can include the license key, with the code below. Please make sure that the file that contains the license key is included in your project and is copied during Copy Bundle Resources build phase.

Swift

MBIMicroblinkSDK.shared().setLicenseResource("license-key-file", withExtension: "txt", inSubdirectory: "directory-to-license-key", for: Bundle.main)

Objective-C

[[MBIMicroblinkSDK sharedInstance] setLicenseResource:@"license-key-file" withExtension:@"txt" inSubdirectory:@"" forBundle:[NSBundle mainBundle]];

If the licence is invalid or expired then the methods above will throw an exception.

5. Registering for scanning events

In the previous step, you instantiated MBIDocumentOverlayViewController object with a delegate object. This object gets notified on certain events in scanning lifecycle. In this example we set it to self. The protocol which the delegate has to implement is MBIDocumentOverlayViewControllerDelegate protocol. It is necessary to conform to that protocol. We will discuss more about protocols in Advanced integration section. You can use the following default implementation of the protocol to get you started.

Swift

func documentOverlayViewControllerDidFinishScanning(_ documentOverlayViewController: MBIDocumentOverlayViewController, state: MBIRecognizerResultState) {

    // this is done on background thread
    // check for valid state
    if state == .valid {

        // first, pause scanning until we process all the results
        documentOverlayViewController.recognizerRunnerViewController?.pauseScanning()

        DispatchQueue.main.async(execute: {() -> Void in
            // All UI interaction needs to be done on main thread
        })
    }
}

func documentOverlayViewControllerDidTapClose(_ documentOverlayViewController: MBIDocumentOverlayViewController) {
    // Your action on cancel
}

Objective-C

- (void)documentOverlayViewControllerDidFinishScanning:(MBIDocumentOverlayViewController *)documentOverlayViewController state:(MBIRecognizerResultState)state {

    // this is done on background thread
    // check for valid state
    if (state == MBIRecognizerResultStateValid) {

        // first, pause scanning until we process all the results
        [documentOverlayViewController.recognizerRunnerViewController pauseScanning];

        dispatch_async(dispatch_get_main_queue(), ^{
            // All UI interaction needs to be done on main thread
        });
    }
}

- (void)documentOverlayViewControllerDidTapClose:(MBIDocumentOverlayViewController *)documentOverlayViewController {
    // Your action on cancel
}

Advanced BlinkInput integration instructions

This section covers more advanced details of BlinkInput integration.

1. First part will cover the possible customizations when using UI provided by the SDK.
2. Second part will describe how to embed MBIRecognizerRunnerViewController's delegates into your UIViewController with the goal of creating a custom UI for scanning, while still using camera management capabilites of the SDK.
3. Third part will describe how to use the MBIRecognizerRunner (Direct API) for recognition directly from UIImage without the need of camera or to recognize camera frames that are obtained by custom camera management.
4. Fourth part will describe recognizer concept and available recognizers.

Built-in overlay view controllers and overlay subviews

Within BlinkInput SDK there are several built-in overlay view controllers and scanning subview overlays that you can use to perform scanning.

Using MBIBarcodeOverlayViewController

MBIBarcodeOverlayViewController is overlay view controller best suited for performing scanning of various barcodes. It has MBIBarcodeOverlayViewControllerDelegate delegate which can be used out-of-the-box to perform scanning using the default UI. Here is an example how to use and initialize MBIBarcodeOverlayViewController:

Swift

/** Create your overlay view controller */
let barcodeOverlayViewController : MBIBarcodeOverlayViewController = MBIBarcodeOverlayViewController(settings: barcodeSettings, recognizerCollection: recognizerCollection, delegate: self)

/** Create recognizer view controller with wanted overlay view controller */
let recognizerRunneViewController : UIViewController = MBIViewControllerFactory.recognizerRunnerViewController(withOverlayViewController: barcodeOverlayViewController)

/** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
self.present(recognizerRunneViewController, animated: true, completion: nil)

Objective-C

MBIBarcodeOverlayViewController *overlayVC = [[MBIBarcodeOverlayViewController alloc] initWithSettings:settings recognizerCollection: recognizerCollection delegate:self];
UIViewController<MBIRecognizerRunnerViewController>* recognizerRunnerViewController = [MBIViewControllerFactory recognizerRunnerViewControllerWithOverlayViewController:overlayVC];

/** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
[self presentViewController:recognizerRunnerViewController animated:YES completion:nil];

As you can see, when initializing MBIBarcodeOverlayViewController, we are sending delegate property as self. To get results, we need to conform to MBIBarcodeOverlayViewControllerDelegate protocol.

Using MBIFieldByFieldOverlayViewController

MBIFieldByFieldOverlayViewController is overlay view controller best suited for performing scanning of various payment slips and barcodes with field of view. It has MBIFieldByFieldOverlayViewControllerDelegate delegate which can be used out-of-the-box to perform scanning using the default UI. Here is an example how to use and initialize MBIFieldByFieldOverlayViewController:

Swift

/** Create your overlay view controller */
let fieldByFieldOverlayViewController : MBIFieldByFieldOverlayViewController = MBIFieldByFieldOverlayViewController(settings: fieldByFieldOverlaySettings, recognizerCollection: recognizerCollection, delegate: self)

/** Create recognizer view controller with wanted overlay view controller */
let recognizerRunneViewController : UIViewController = MBIViewControllerFactory.recognizerRunnerViewController(withOverlayViewController: fieldByFieldOverlayViewController)

/** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
self.present(recognizerRunneViewController, animated: true, completion: nil)

Objective-C

MBIFieldByFieldOverlayViewController *overlayVC = [[MBIFieldByFieldOverlayViewController alloc] initWithSettings:settings recognizerCollection: recognizerCollection delegate:self];
UIViewController<MBIRecognizerRunnerViewController>* recognizerRunnerViewController = [MBIViewControllerFactory recognizerRunnerViewControllerWithOverlayViewController:overlayVC];

/** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
[self presentViewController:recognizerRunnerViewController animated:YES completion:nil];

As you can see, when initializing MBIFieldByFieldOverlayViewController, we are sending delegate property as self. To get results, we need to conform to MBIFieldByFieldOverlayViewControllerDelegate protocol.

Using MBIDocumentCaptureOverlayViewController

MBIDocumentCaptureOverlayViewController is overlay view controller best suited for performing captureing cropped document images. It has MBIDocumentCaptureOverlayViewControllerDelegate delegate which can be used out-of-the-box to perform scanning using the default UI. Here is an example how to use and initialize MBIDocumentCaptureOverlayViewController:

Swift

/** Create your overlay view controller */
let documentCaptureOverlayViewController : MBIDocumentCaptureOverlayViewController = MBIDocumentCaptureOverlayViewController(settings: settings, recognizer: documentCaptureRecognizer, delegate: self)

/** Create recognizer view controller with wanted overlay view controller */
let recognizerRunneViewController : UIViewController = MBIViewControllerFactory.recognizerRunnerViewController(withOverlayViewController: documentCaptureOverlayViewController)

/** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
self.present(recognizerRunneViewController, animated: true, completion: nil)

Objective-C

MBIDocumentCaptureOverlayViewController *overlayVC = [[MBIDocumentCaptureOverlayViewController alloc] initWithSettings:settings recognizer: documentCaptureRecognizer delegate:self];
UIViewController<MBIRecognizerRunnerViewController>* recognizerRunnerViewController = [MBIViewControllerFactory recognizerRunnerViewControllerWithOverlayViewController:overlayVC];

/** Present the recognizer runner view controller. You can use other presentation methods as well (instead of presentViewController) */
[self presentViewController:recognizerRunnerViewController animated:YES completion:nil];

As you can see, when initializing MBIDocumentCaptureOverlayViewController, we are sending delegate property as self. To get results, we need to conform to MBIDocumentCaptureOverlayViewControllerDelegate protocol.

Custom overlay view controller

Please check our Samples for custom implementation of overlay view controller.

Overlay View Controller is an abstract class for all overlay views.

It’s responsibility is to provide meaningful and useful interface for the user to interact with.

Typical actions which need to be allowed to the user are:

• intuitive and meaniningful way to guide the user through scanning process. This is usually done by presenting a “viewfinder” in which the user need to place the scanned object
• a way to cancel the scanning, typically with a “cancel” or “back” button
• a way to power on and off the light (i.e. “torch”) button

BlinkInput SDK always provides it’s own default implementation of the Overlay View Controller for every specific use. Your implementation should closely mimic the default implementation as it’s the result of thorough testing with end users. Also, it closely matches the underlying scanning technology.

For example, the scanning technology usually gives results very fast after the user places the device’s camera in the expected way above the scanned object. This means a progress bar for the scan is not particularly useful to the user. The majority of time the user spends on positioning the device’s camera correctly. That’s just an example which demonstrates careful decision making behind default camera overlay view.

1. Subclassing

To use your custom overlay with Microblink’s camera view, you must first subclass MBICustomOverlayViewController and implement the overlay behaviour conforming wanted protocols.

2. Protocols

There are five MBIRecognizerRunnerViewController protocols and one overlay protocol MBIOverlayViewControllerInterface.

Five RecognizerRunnerView protocols are:

• MBIScanningRecognizerRunnerViewControllerDelegate
• MBIDetectionRecognizerRunnerViewControllerDelegate
• MBIOcrRecognizerRunnerViewControllerDelegate
• MBIDebugRecognizerRunnerViewControllerDelegate
• MBIRecognizerRunnerViewControllerDelegate

In viewDidLoad, other protocol conformation can be done and it’s done on recognizerRunnerViewController property of MBIOverlayViewController, for example:

Swift and Objective-C

self.scanningRecognizerRunnerViewControllerDelegate = self;

3. Initialization

In Quick Start guide it is shown how to use a default overlay view controller. You can now swap default view controller with your implementation of CustomOverlayViewController

Swift

let recognizerRunnerViewController : UIViewController = MBIViewControllerFactory.recognizerRunnerViewController(withOverlayViewController: CustomOverlayViewController)

Objective-C

UIViewController<MBIRecognizerRunnerViewController>* recognizerRunnerViewController = [MBIViewControllerFactory recognizerRunnerViewControllerWithOverlayViewController:CustomOverlayViewController];

Direct processing API

This guide will in short present you how to process UIImage objects with BlinkInput SDK, without starting the camera video capture.

With this feature you can solve various use cases like: - recognizing text on images in Camera roll - taking full resolution photo and sending it to processing - scanning barcodes on images in e-mail etc.

DirectAPI-sample demo app here will present UIImagePickerController for taking full resolution photos, and then process it with BlinkInput SDK to get scanning results using Direct processing API.

Direct processing API is handled with MBIRecognizerRunner. That is a class that handles processing of images. It also has protocols as MBIRecognizerRunnerViewController. Developer can choose which protocol to conform:

• MBIScanningRecognizerRunnerDelegate
• MBIDetectionRecognizerRunnerDelegate
• MBIDebugRecognizerRunnerDelegate
• MBIOcrRecognizerRunnerDelegate

In example, we are conforming to MBIScanningRecognizerRunnerDelegate protocol.

To initiate the scanning process, first decide where in your app you want to add scanning functionality. Usually, users of the scanning library have a button which, when tapped, starts the scanning process. Initialization code is then placed in touch handler for that button. Here we’re listing the initialization code as it looks in a touch handler method.

Swift

func setupRecognizerRunner() {
    var recognizers = [MBIRecognizer]()
    pdf417Recognizer = MBIPdf417Recognizer()
    recognizers.append(pdf417Recognizer!)
    let recognizerCollection = MBIRecognizerCollection(recognizers: recognizers)
    recognizerRunner = MBIRecognizerRunner(recognizerCollection: recognizerCollection)
    recognizerRunner?.scanningRecognizerRunnerDelegate = self
}

func processImageRunner(_ originalImage: UIImage) {
    var image: MBIImage? = nil
    if let anImage = originalImage {
        image = MBIImage(uiImage: anImage)
    }
    image?.cameraFrame = true
    image?.orientation = MBIProcessingOrientation.left
    let _serialQueue = DispatchQueue(label: "com.microblink.DirectAPI-sample-swift")
    _serialQueue.async(execute: {() -> Void in
        self.recognizerRunner?.processImage(image!)
    })
}

func recognizerRunner(_ recognizerRunner: MBIRecognizerRunner, didFinishScanningWith state: MBIRecognizerResultState) {
    if blinkInputRecognizer.result.resultState == MBIRecognizerResultStateValid {
        // Handle result
    }
}

Objective-C

- (void)setupRecognizerRunner {
    NSMutableArray<MBIRecognizer *> *recognizers = [[NSMutableArray alloc] init];

    self.pdf417Recognizer = [[MBIPdf417Recognizer alloc] init];

    [recognizers addObject: self.pdf417Recognizer];

    MBIRecognizerCollection *recognizerCollection = [[MBIRecognizerCollection alloc] initWithRecognizers:recognizers];

    self.recognizerRunner = [[MBIRecognizerRunner alloc] initWithRecognizerCollection:recognizerCollection];
    self.recognizerRunner.scanningRecognizerRunnerDelegate = self;
}

- (void)processImageRunner:(UIImage *)originalImage {
    MBIImage *image = [MBIImage imageWithUIImage:originalImage];
    image.cameraFrame = YES;
    image.orientation = MBIProcessingOrientationLeft;
    dispatch_queue_t _serialQueue = dispatch_queue_create("com.microblink.DirectAPI-sample", DISPATCH_QUEUE_SERIAL);
    dispatch_async(_serialQueue, ^{
        [self.recognizerRunner processImage:image];
    });
}

- (void)recognizerRunner:(nonnull MBIRecognizerRunner *)recognizerRunner didFinishScanningWithState:(MBIRecognizerResultState)state {
    if (self.blinkInputRecognizer.result.resultState == MBIRecognizerResultStateValid) {
        // Handle result
    }
}

Now you’ve seen how to implement the Direct processing API.

In essence, this API consists of two steps:

• Initialization of the scanner.
• Call of - (void)processImage:(MBIImage *)image; method for each UIImage or CMSampleBufferRef you have.

Using Direct API for NSString recognition (parsing)

Some recognizers support recognition from NSString. They can be used through Direct API to parse given NSString and return data just like when they are used on an input image. When recognition is performed on NSString, there is no need for the OCR. Input NSString is used in the same way as the OCR output is used when image is being recognized. Recognition from String can be performed in the same way as recognition from image. The only difference is that user should call - (void)processString:(NSString *)string; on MBIRecognizerRunner.

MBIRecognizer and available recognizers

The MBIRecognizer concept

The MBIRecognizer is the basic unit of processing within the SDK. Its main purpose is to process the image and extract meaningful information from it. As you will see later, the SDK has lots of different MBIRecognizer objects that have various purposes.

Each MBIRecognizer has a MBIRecognizerResult object, which contains the data that was extracted from the image. The MBIRecognizerResult object is a member of corresponding MBIRecognizer object its lifetime is bound to the lifetime of its parent MBIRecognizer object. If you need your MBIRecognizerResult object to outlive its parent MBIRecognizer object, you must make a copy of it by calling its method copy.

While MBIRecognizer object works, it changes its internal state and its result. The MBIRecognizer object’s MBIRecognizerResult always starts in Empty state. When corresponding MBIRecognizer object performs the recognition of given image, its MBIRecognizerResult can either stay in Empty state (in case MBIRecognizerfailed to perform recognition), move to Uncertain state (in case MBIRecognizer performed the recognition, but not all mandatory information was extracted) or move to Valid state (in case MBIRecognizer performed recognition and all mandatory information was successfully extracted from the image).

As soon as one MBIRecognizer object’s MBIRecognizerResult within MBIRecognizerCollection given to MBIRecognizerRunner or MBIRecognizerRunnerViewController changes to Valid state, the onScanningFinished callback will be invoked on same thread that performs the background processing and you will have the opportunity to inspect each of your MBIRecognizer objects’ MBIRecognizerResult to see which one has moved to Valid state.

As soon as onScanningFinished method ends, the MBIRecognizerRunnerViewController will continue processing new camera frames with same MBIRecognizer objects, unless paused. Continuation of processing or reset recognition will modify or reset all MBIRecognizer objects’s MBIRecognizerResult. When using built-in activities, as soon as onScanningFinished is invoked, built-in activity pauses the MBIRecognizerRunnerViewController and starts finishing the activity, while saving the MBIRecognizerCollection with active MBIRecognizer.

MBIRecognizerCollection concept

The MBIRecognizerCollection is is wrapper around MBIRecognizer objects that has array of MBIRecognizer objects that can be used to give MBIRecognizer objects to MBIRecognizerRunner or MBIRecognizerRunnerViewController for processing.

The MBIRecognizerCollection is always constructed with array [[MBIRecognizerCollection alloc] initWithRecognizers:recognizers] of MBIRecognizer objects that need to be prepared for recognition (i.e. their properties must be tweaked already).

The MBIRecognizerCollection manages a chain of MBIRecognizer objects within the recognition process. When a new image arrives, it is processed by the first MBIRecognizer in chain, then by the second and so on, iterating until a MBIRecognizer object’s MBIRecognizerResult changes its state to Valid or all of the MBIRecognizer objects in chain were invoked (none getting a Valid result state).

You cannot change the order of the MBIRecognizer objects within the chain - no matter the order in which you give MBIRecognizer objects to MBIRecognizerCollection, they are internally ordered in a way that provides best possible performance and accuracy. Also, in order for SDK to be able to order MBIRecognizer objects in recognition chain in a best way possible, it is not allowed to have multiple instances of MBIRecognizer objects of the same type within the chain. Attempting to do so will crash your application.

List of available recognizers

This section will give a list of all MBIRecognizer objects that are available within BlinkInput SDK, their purpose and recommendations how they should be used to get best performance and user experience.

Frame Grabber Recognizer

The MBIFrameGrabberRecognizer is the simplest recognizer in SDK, as it does not perform any processing on the given image, instead it just returns that image back to its onFrameAvailable. Its result never changes state from empty.

This recognizer is best for easy capturing of camera frames with MBIRecognizerRunnerViewController. Note that MBIImage sent to onFrameAvailable are temporary and their internal buffers all valid only until the onFrameAvailable method is executing - as soon as method ends, all internal buffers of MBIImage object are disposed. If you need to store MBIImage object for later use, you must create a copy of it by calling copy.

Success Frame Grabber Recognizer

The MBISuccessFrameGrabberRecognizer is a special MBIecognizer that wraps some other MBIRecognizer and impersonates it while processing the image. However, when the MBIRecognizer being impersonated changes its MBIRecognizerResult into Valid state, the MBISuccessFrameGrabberRecognizer captures the image and saves it into its own MBISuccessFrameGrabberRecognizerResult object.

Since MBISuccessFrameGrabberRecognizer impersonates its slave MBIRecognizer object, it is not possible to give both concrete MBIRecognizer object and MBISuccessFrameGrabberRecognizer that wraps it to same MBIRecognizerCollection - doing so will have the same result as if you have given two instances of same MBIRecognizer type to the MBIRecognizerCollection - it will crash your application.

This recognizer is best for use cases when you need to capture the exact image that was being processed by some other MBIRecognizer object at the time its MBIRecognizerResult became Valid. When that happens, MBISuccessFrameGrabberRecognizer's MBISuccessFrameGrabberRecognizerResult will also become Valid and will contain described image.

PDF417 recognizer

The MBIPdf417Recognizer is recognizer specialised for scanning PDF417 2D barcodes. This recognizer can recognize only PDF417 2D barcodes - for recognition of other barcodes, please refer to BarcodeRecognizer.

This recognizer can be used in any overlay view controller, but it works best with the MBIBarcodeOverlayViewController, which has UI best suited for barcode scanning.

Barcode recognizer

The MBIBarcodeRecognizer is recognizer specialised for scanning various types of barcodes. This recognizer should be your first choice when scanning barcodes as it supports lots of barcode symbologies, including the PDF417 2D barcodes, thus making PDF417 recognizer possibly redundant, which was kept only for its simplicity.

You can enable multiple barcode symbologies within this recognizer, however keep in mind that enabling more barcode symbologies affect scanning performance - the more barcode symbologies are enabled, the slower the overall recognition performance. Also, keep in mind that some simple barcode symbologies that lack proper redundancy, such as Code 39, can be recognized within more complex barcodes, especially 2D barcodes, like PDF417.

This recognizer can be used in any overlay view controller, but it works best with the MBIBarcodeOverlayViewController, which has UI best suited for barcode scanning.

BlinkInput recognizer

The MBIBlinkInputRecognizer is generic OCR recognizer used for scanning segments which enables specifying MBIProcessors that will be used for scanning. Most commonly used MBIProcessor within this recognizer is MBIParserGroupProcessor) that activates all MBIParsers in the group to extract data of interest from the OCR result.

This recognizer can be used in any context. It is used internally in the implementation of the provided MBIFieldByFieldOverlayViewController.

MBIProcessors are explained in The Processor concept section and you can find more about MBIParsers in The Parser concept section.

Detector recognizer

The MBIDetectorRecognizer is recognizer for scanning generic documents using custom MBIDetector. You can find more about Detector in The Detector concept section. MBIDetectorRecognizer can be used simply for document detection and obtaining its image. The more interesting use case is data extraction from the custom document type. MBIDetectorRecognizer performs document detection and can be configured to extract fields of interest from the scanned document by using Templating API. You can find more about Templating API in this section.

Document Capture recognizer

The MBIDocumentCaptureRecognizer is used for taking cropped document images. This recognizer can be used in any context, but it works best with the MBIDocumentCaptureOverlayViewController which takes high resolution document images and guides the user through the image capture process.

MBIProcessor and MBIParser

The MBIProcessors and MBIParsers are standard processing units within BlinkInput SDK used for data extraction from the input images. Unlike the MBIRecognizer, MBIProcessor and MBIParser are not stand-alone processing units. MBIProcessor is always used within MBIRecognizer and MBIParser is used within appropriate MBIProcessor to extract data from the OCR result.

The MBIProcessor concept

MBIProcessor is a processing unit used within some Recognizer which supports processors. It process the input image prepared by the enclosing Recognizer in the way that is characteristic to the implementation of the concrete MBIProcessor.

MBIProcessor architecture is similar to MBIRecognizer architecture described in The Recognizer concept section. Each instance also has associated inner MBIRecognizerResult object whose lifetime is bound to the lifetime of its parent MBIProcessor object and it is updated while MBIProcessor works. If you need your MBIRecognizerResult object to outlive its parent MBIProcessor object, you must make a copy of it by calling its method copy.

It also has its internal state and while it is in the working state during recognition process, it is not allowed to tweak MBIProcessor object’s properties.

To support common use cases, there are several different MBIProcessor implementations available. They are listed in the next section.

List of available processors

This section will give a list of MBIProcessor types that are available within BlinkInput SDK and their purpose.

Image Return Processor

The MBIImageReturnProcessor is used for obtaining input images. It simply saves the input image and makes it available after the scanning is done.

The appearance of the input image depends on the context in which MBIImageReturnProcessor is used. For example, when it is used within MBIBlinkInputRecognizer, simply the raw image of the scanning region is processed. When it is used within the Templating API, input image is dewarped (cropped and rotated).

The image is returned as the raw MBIImage type. Also, processor can be configured to encode saved image to JPEG.

Parser Group Processor

The MBIParserGroupProcessor is the type of the processor that performs the OCR (Optical Character Recognition) on the input image and lets all the parsers within the group to extract data from the OCR result. The concept of MBIParser is described in the next section.

Before performing the OCR, the best possible OCR engine options are calculated by combining engine options needed by each MBIParser from the group. For example, if one parser expects and produces result from uppercase characters and other parser extracts data from digits, both uppercase characters and digits must be added to the list of allowed characters that can appear in the OCR result. This is a simplified explanation because OCR engine options contain many parameters which are combined by the MBIParserGroupProcessor.

Because of that, if multiple parsers and multiple parser group processors are used during the scan, it is very important to group parsers carefully.

Let’s see this on an example: assume that we have two parsers at our disposal: MBIAmountParser and MBIEmailParser. MBIAmountParser knows how to extract amount’s from OCR result and requires from OCR only to recognize digits, periods and commas and ignore letters. On the other hand, MBIEmailParser knows how to extract e-mails from OCR result and requires from OCR to recognize letters, digits, ‘@’ characters and periods, but not commas.

If we put both MBIAmountParser and MBIEmailParser into the same MBIParserGroupProcessor, the merged OCR engine settings will require recognition of all letters, all digits, ‘@’ character, both period and comma. Such OCR result will contain all characters for MBIEmailParser to properly parse e-mail, but might confuse MBIAmountParser if OCR misclassifies some characters into digits.

If we put MBIAmountParser in one MBIParserGroupProcessor and MBIEmailParser in another MBIParserGroupProcessor, OCR will be performed for each parser group independently, thus preventing the MBIAmountParser confusion, but two OCR passes of the image will be performed, which can have a performance impact.

MBIParserGroupProcessor is most commonly used MBIProcessor. It is used whenever the OCR is needed. After the OCR is performed and all parsers are run, parsed results can be obtained through parser objects that are enclosed in the group. MBIParserGroupProcessor instance also has associated inner MBIParserGroupProcessorResult whose state is updated during processing and its property ocrLayout can be used to obtain the raw MBIOcrLayout that was used for parsing data.

Take note that MBIOcrLayout is available only if it is allowed by the BlinkInput SDK license key. MBIOcrLayout structure contains information about all recognized characters and their positions on the image. To prevent someone to abuse that, obtaining of the MBIOcrLayout structure is allowed only by the premium license keys.

The MBIParser concept

MBIParser is a class of objects that are used to extract structured data from the raw OCR result. It must be used within MBIParserGroupProcessor which is responsible for performing the OCR, so MBIParser is not stand-alone processing unit.

Like MBIRecognizer and all other processing units, each MBIParser instance has associated inner MBIRecognizerResult object whose lifetime is bound to the lifetime of its parent MBIParser object and it is updated while MBIParser works. When parsing is done MBIParserResult can be used for obtaining extracted data. If you need your MBIParserResult object to outlive its parent MBIParser object, you must make a copy of it by calling its method copy.

It also has its internal state and while it is in the working state during recognition process, it is not allowed to tweak MBIParser object’s properties.

There are a lot of different MBIParsers for extracting most common fields which appear on various documents. Also, most of them can be adjusted for specific use cases. For all other custom data fields, there is RegexParser available which can be configured with the arbitrary regular expression.

List of available parsers

Amount Parser

MBIAmountParser is used for extracting amounts from the OCR result.

Date Parser

MBIDateParser is used for extracting dates in various formats from the OCR result.

Email Parser

MBIEmailParser is used for extracting e-mail addresses from the OCR result.

IBAN Parser

MBIIbanParser is used for extracting IBAN (International Bank Account Number) from the OCR result.

License Plates Parser

MBILicensePlatesParser is used for extracting license plate content from the OCR result.

Raw Parser

MBIRawParser is used for obtaining string version of raw OCR result, without performing any smart parsing operations.

Regex Parser

MBIRegexParser is used for extracting OCR result content which is in accordance with the given regular expression. Regular expression parsing is not performed with java’s regex engine. Instead, it is performed with custom regular expression engine.

TopUp Parser

MBITopUpParser is used for extracting TopUp (mobile phone coupon) codes from the OCR result. There exists TopUpPreset enum with presets for most common vendors. Method - (void)setTopUpPreset:(MBITopUpPreset)topUpPreset can be used to configure parser to only return codes with the appropriate format defined by the used preset.

VIN (Vehicle Identification Number) Parser

MBIVinParser is used for extracting VIN (Vehicle Identification Number) from the OCR result.

Scanning generic documents with Templating API

This section discusses the setting up of MBIDetectorRecognizer for scanning templated documents. Please check Templating-sample sample app for source code examples.

Templated document is any document which is defined by its template. Template contains the information about how the document should be detected, i.e. found on the camera scene and information about which part of the document contains which useful information.

Defining how document should be detected

Before performing OCR of the document, BlinkInput first needs to find its location on a camera scene. In order to perform detection, you need to define MBIDetector.

You have to set concrete MBIDetector when instantiating the MBIDetectorRecognizer as a parameter to its constructor.

You can find out more information about detectors that can be used in section List of available detectors. The most commonly used detector is MBIDocumentDetector.

Defining how fields of interest should be extracted

MBIDetector produces its result which contains document location. After the document has been detected, all further processing is done on the detected part of the input image.

There may be one or more variants of the same document type, for example for some document there may be old and new version and both of them must be supported. Because of that, for implementing support for each document, one or multiple templating classes are used. MBITemplatingClass is described in The Templating Class component section.

MBITemplatingClass holds all needed information and components for processing its class of documents. Templating classes are processed in chain, one by one. On first class for which the data is successfully extracted, the chain is terminated and recognition results are returned. For each input image processing is done in the following way:

1. Classification MBIProcessorGroups are run on the defined locations to extract data. MBIProcessorGroup is used to define the location of interest on the detected document and MBIProcessors that will extract data from that location. You can find more about MBIProcessorGroup in the next section.

2. MBITemplatingClassifier is run with the data extracted by the classification processor groups to decide whether the currently scanned document belongs to the current class or not. Its classify method simply returns YES/true or NO/false. If the classifier returns NO/false, recognition is moved to the next class in the chain, if it exists. You can find more about MBITemplatingClassifier in this section.

3. If the MBITemplatingClassifier has decided that currently scanned document belongs to the current class, non-classification MBIProcessorGroups are run to extract other fields of interest.

The MBIProcessorGroup component

In templating API MBIProcessorGroup is used to define the location of the field of interest on the detected document and how that location should be processed by setting following parameters in its constructor:

1. Location coordinates relative to document detection which are passed as [Rectangle] object.

2. MBIDewarpPolicy which determines the resulting image chunk for processing. You can find a description of each MBIDewarpPolicy, its purpose and recommendations when it should be used to get the best results in List of available dewarp policies section.

3. Collection of processors that will be executed on the prepared chunk of the image for current document location. You can find more information about processors in The Processor concept section.

List of available dewarp policies

Concrete MBIDewarpPolicy defines how specific location of interest should be dewarped (cropped and rotated). It determines the height and width of the resulting dewarped image in pixels. Here is the list of available dewarp policies with linked doc for more information:

• MBIFixedDewarpPolicy

  • defines the exact height of the dewarped image in pixels
  • usually the best policy for processor groups that use a legacy OCR engine

• MBIDPIBasedDewarpPolicy:

  • defines the desired DPI (Dots Per Inch)
  • the height of the dewarped image will be calculated based on the actual physical size of the document provided by the used detector and chosen DPI
  • usually the best policy for processor groups that prepare location’s raw image for output

• MBINoUpScalingDewarpPolicy:

  • defines the maximal allowed height of the dewarped image in pixels
  • the height of the dewarped image will be calculated in a way that no part of the image will be up-scaled
  • if the height of the resulting image is larger than maximal allowed, then the maximal allowed height will be used as actual height, which effectively scales down the image
  • usually the best policy for processors that use neural networks, for example, DEEP OCR, hologram detection or NN-based classification

The MBITemplatingClass component

MBITemplatingClass enables implementing support for a specific class of documents that should be scanned with templating API. Final implementation of the templating recognizer consists of one or more templating classes, one class for each version of the document.

MBITemplatingClass contains two collections of MBIProcessorGroups and a MBITemplatingClassifier.

The two collections of processor groups within MBITemplatingClass are:

1. The classification processor groups which are set by using the [- (void)setClassificationProcessorGroups:(nonnull NSArray<__kindof MBIProcessorGroup *> *)processorGroups] method. MBIProcessorGroups from this collection will be executed before classification, which means that they are always executed when processing comes to this class.

2. The non-classification processor groups which are set by using the [- (void)setNonClassificationProcessorGroups:(nonnull NSArray<__kindof MBIProcessorGroup *> *)processorGroups]method. MBIProcessorGroups from this collection will be executed after classification if the classification has been positive.

A component which decides whether the scanned document belongs to the current class is MBITemplatingClass. It can be set by using the - (void)setTemplatingClassifier:(nullable id<MBITemplatingClassifier>)templatingClassifier method. If it is not set, non-classification processor groups will not be executed. Instructions for implementing the MBITemplatingClassifier are given in the next section.

Implementing the MBITemplatingClassifier

Each concrete templating classifier implements the MBITemplatingClassifier interface, which requires to implement its classify method that is invoked while evaluating associated MBITemplatingClass.

Classification decision should be made based on the processing result which is returned by one or more processing units contained in the collection of the classification processor groups. As described in The ProcessorGroup component section, each processor group contains one or more MBIProcessors. There are different MBIProcessors which may enclose smaller processing units, for example, MBIParserGroupProcessor maintains the group of MBIParsers. Result from each of the processing units in that hierarchy can be used for classification. In most cases MBIParser result is used to determine whether some data in the expected format exists on the specified location.

To be able to retrieve results from the various processing units that are needed for classification, their instances must be available when classify method is called.

Obtaining recognition results

When recognition is done, results can be obtained through processing units instances, such as: MBIProcessors, MBIParsers, etc. which are used for configuring the MBITemplatingRecognizer and later for processing the input image.

The MBIDetector concept

MBIDetector is a processing unit used within some MBIRecognizer which supports detectors, such as MBIDetectorRecognizer. Concrete MBIDetector knows how to find the certain object on the input image. MBIRecognizer can use it to perform object detection prior to performing further recognition of detected object’s contents.

MBIDetector architecture is similar to MBIRecognizer architecture described in The Recognizer concept section. Each instance also has associated inner MBIRecognizerResult object whose lifetime is bound to the lifetime of its parent MBIDetector object and it is updated while MBIDetector works. If you need your MBIRecognizerResult object to outlive its parent MBIDetector object, you must make a copy of it by calling its copy method.

It also has its internal state and while it is in the working state during recognition process, it is not allowed to tweak MBIDetector object’s properties.

When detection is performed on the input image, each MBIDetector in its associated MBIDetectorResult object holds the following information:

• MBIDetectionCode that indicates the type of the detection.

• MBIDetectionStatus that represents the status of the detection.

• each concrete detector returns additional information specific to the detector type

To support common use cases, there are several different MBIDetector implementations available. They are listed in the next section.

List of available detectors

Document Detector

MBIDocumentDetector is used to detect card documents, cheques, A4-sized documents, receipts and much more.

It accepts one or more MBIDocumentSpecifications. MBIDocumentSpecification represents a specification of the document that should be detected by using edge detection algorithm and predefined aspect ratio.

For the most commonly used document formats, there is a helper method + (instancetype)createFromPreset:(MBIDocumentSpecificationPreset)preset which creates and initializes the document specification based on the given MBIDocumentSpecificationPreset.

For the list of all available configuration methods see MBIDocumentDetector doc, and for available result content see MBIDocumentDetectorResult doc.

MRTD Detector

MBIMrtdDetector is used to perform detection of Machine Readable Travel Documents (MRTD).

Method - (void)setMrtdSpecifications:(NSArray<__kindof MBIMrtdSpecification *> *)mrtdSpecifications can be used to define which MRTD documents should be detectable. It accepts the array of MBIMrtdSpecification. MBIMrtdSpecification represents specification of MRTD that should be detected. It can be created from the MBIMrtdSpecificationPreset by using + (instancetype)createFromPreset:(MBIMrtdSpecificationPreset)preset method.

If MBIMrtdSpecifications are not set, all supported MRTD formats will be detectable.

For the list of all available configuration methods see MBIMrtdDetector doc, and for available result content see MBIMrtdDetectorResult doc.

Localization

The SDK is localized on following languages: Arabic, Chinese simplified, Chinese traditional, Croatian, Czech, Dutch, Filipino, French, German, Hebrew, Hungarian, Indonesian, Italian, Malay, Portuguese, Romanian, Slovak, Slovenian, Spanish, Thai, Vietnamese.

If you would like us to support additional languages or report incorrect translation, please contact us at help.microblink.com.

If you want to add additional languages yourself or change existing translations, you need to set customLocalizationFileName property on MBIMicroblinkApp object to your strings file name.

For example, let’s say that we want to change text “Scan the front side of a document” to “Scan the front side” in BlinkID sample project. This would be the steps:

• Find the translation key in en.strings file inside BlinkInput.framework
• Add a new file MyTranslations.strings to the project by using “Strings File” template
• With MyTranslations.string open, in File inspector tap “Localize…” button and select English
• Add the translation key “blinkid_generic_message” and the value “Scan the front side” to MyTranslations.strings
• Finally in AppDelegate.swift in method application(_:, didFinishLaunchingWithOptions:) add MBIMicroblinkApp.instance()?.customLocalizationFileName = "MyTranslations"

Troubleshooting

Integration problems

In case of problems with integration of the SDK, first make sure that you have tried integrating it into XCode by following integration instructions.

If you have followed XCode integration instructions and are still having integration problems, please contact us at help.microblink.com.

SDK problems

In case of problems with using the SDK, you should do as follows:

Licencing problems

If you are getting “invalid licence key” error or having other licence-related problems (e.g. some feature is not enabled that should be or there is a watermark on top of camera), first check the console. All licence-related problems are logged to error log so it is easy to determine what went wrong.

When you have determine what is the licence-relate problem or you simply do not understand the log, you should contact us help.microblink.com. When contacting us, please make sure you provide following information:

• exact Bundle ID of your app (from your info.plist file)
• licence that is causing problems
• please stress out that you are reporting problem related to iOS version of BlinkInput SDK
• if unsure about the problem, you should also provide excerpt from console containing licence error

Other problems

If you are having problems with scanning certain items, undesired behaviour on specific device(s), crashes inside BlinkInput SDK or anything unmentioned, please do as follows:

• Contact us at help.microblink.com describing your problem and provide following information:
  • log file obtained in previous step
  • high resolution scan/photo of the item that you are trying to scan
  • information about device that you are using
  • please stress out that you are reporting problem related to iOS version of BlinkInput SDK

Frequently asked questions and known problems

Here is a list of frequently asked questions and solutions for them and also a list of known problems in the SDK and how to work around them.

Note on ARM Macs

We are supporting ARM64 Device slice through our .xcframework format. We are still in development supporting ARM64 Simulator slice for newly released ARM Macs and we will update our SDK with ARM64 Simulator support as soon as development is done.

In demo everything worked, but after switching to production license I get NSError with MBIMicroblinkSDKRecognizerErrorDomain and MBIRecognizerFailedToInitalize code as soon as I construct specific MBIRecognizer object

Each license key contains information about which features are allowed to use and which are not. This NSError indicates that your production license does not allow using of specific MBIRecognizer object. You should contact support to check if provided licence is OK and that it really contains all features that you have purchased.

I get NSError with MBIMicroblinkSDKRecognizerErrorDomain and MBIRecognizerFailedToInitalize code with trial license key

Whenever you construct any MBIRecognizer object or, a check whether license allows using that object will be performed. If license is not set prior constructing that object, you will get NSError with MBIMicroblinkSDKRecognizerErrorDomain and MBIRecognizerFailedToInitalize code. We recommend setting license as early as possible in your app.

Undefined Symbols on Architecture armv7

Make sure you link your app with iconv and Accelerate frameworks as shown in Quick start. If you are using Cocoapods, please be sure that you’ve installed git-lfs prior to installing pods. If you are still getting this error, go to project folder and execute command git-lfs pull.

Crash on armv7 devices

SDK crashes on armv7 devices if bitcode is enabled. We are working on it.

In my didFinish callback I have the result inside my MBIRecognizer, but when scanning activity finishes, the result is gone

This usually happens when using MBIRecognizerRunnerViewController and forgetting to pause the MBIRecognizerRunnerViewController in your didFinish callback. Then, as soon as didFinish happens, the result is mutated or reset by additional processing that MBIRecognizer performs in the time between end of your didFinish callback and actual finishing of the scanning activity. For more information about statefulness of the MBIRecognizer objects, check this section.

Unsupported architectures when submitting app to App Store

BlinkInput.framework is a dynamic framework which contains slices for all architectures - device and simulator. If you intend to extract .ipa file for ad hoc distribution, you’ll need to preprocess the framework to remove simulator architectures.

Ideal solution is to add a build phase after embed frameworks build phase, which strips unused slices from embedded frameworks.

Build step is based on the one provided here: http://ikennd.ac/blog/2015/02/stripping-unwanted-architectures-from-dynamic-libraries-in-xcode/

APP_PATH="${TARGET_BUILD_DIR}/${WRAPPER_NAME}"

# This script loops through the frameworks embedded in the application and
# removes unused architectures.
find "$APP_PATH" -name '*.framework' -type d | while read -r FRAMEWORK
do
FRAMEWORK_EXECUTABLE_NAME=$(defaults read "$FRAMEWORK/Info.plist" CFBundleExecutable)
FRAMEWORK_EXECUTABLE_PATH="$FRAMEWORK/$FRAMEWORK_EXECUTABLE_NAME"
echo "Executable is $FRAMEWORK_EXECUTABLE_PATH"

EXTRACTED_ARCHS=()

for ARCH in $ARCHS
do
echo "Extracting $ARCH from $FRAMEWORK_EXECUTABLE_NAME"
lipo -extract "$ARCH" "$FRAMEWORK_EXECUTABLE_PATH" -o "$FRAMEWORK_EXECUTABLE_PATH-$ARCH"
EXTRACTED_ARCHS+=("$FRAMEWORK_EXECUTABLE_PATH-$ARCH")
done

echo "Merging extracted architectures: ${ARCHS}"
lipo -o "$FRAMEWORK_EXECUTABLE_PATH-merged" -create "${EXTRACTED_ARCHS[@]}"
rm "${EXTRACTED_ARCHS[@]}"

echo "Replacing original executable with thinned version"
rm "$FRAMEWORK_EXECUTABLE_PATH"
mv "$FRAMEWORK_EXECUTABLE_PATH-merged" "$FRAMEWORK_EXECUTABLE_PATH"

done

Disable logging

Logging can be disabled by calling disableMicroblinkLogging method on MBILogger instance.

Size Report

We are delivering complete size report of our BlinkInput SDK based on our BlinkInput-sample-Swift sample project. You can check that here.

Additional info

Complete API reference can be found here.

For any other questions, feel free to contact us at help.microblink.com.