@microblink/blinkid-in-browser-sdk

BlinkID In-browser SDK

BlinkID In-browser SDK enables scanning of various identity documents including driving licenses, national identity cards, passports and others. SDK provides real-time in-browser data extraction, without any need for sending images to servers for processing.

Using BlinkID in your web app requires a valid license key. You can obtain a trial license key by registering to Microblink dashboard. After registering, you will be able to generate a license key for your web app. The license key is bound to fully qualified domain name of your web app, so please make sure you enter the correct name when asked. Also, keep in mind that if you plan to serve your web app from different domains, you will need different license keys.

For more information on how to integrate BlinkID SDK into your web app read the instructions below. Make sure you read the latest Release notes for most recent changes and improvements. For a quick demo, check out our integration sample app. You can also check the live version of this sample app by clicking this link.

BlinkID In-browser SDK is meant to be used natively in a web browser. It will not work correctly within a iOS/Android WebView or NodeJS backend service. If you are looking for Cordova/PhoneGap version, please go here. If you want to use BlinkID as a backend service, please go here.

Table of contents

• BlinkID integration instructions
  • Adding BlinkID as a dependency of your project.
  • Performing your first scan
  • Recognizing still images
  • Optimal deployment of your web app
• The Recognizer concept, RecognizerRunner and VideoRecognizer
  • The Recognizer concept
  • RecognizerRunner
  • Performing recognition of video streams using VideoRecognizer
  • Custom UX with VideoRecognizer
• Handling processing events with MetadataCallbacks
• List of available recognizers
  • Success Frame Grabber Recognizer
  • ID barcode recognizer
  • Machine Readable Travel Document recognizer
• Troubleshooting
  • Integration problems
  • SDK problems
    • Licensing problems
    • Other problems
• FAQ and known issues
• Additional info

BlinkID integration instructions

This repository contains WebAssembly file and it's support JS files which contains the core implementation of BlinkID functionalities. In order to make integration of the WebAssembly easier and more developer friendly, a TypeScript support code is also provided, giving a type-safe and easy to use integration API to the developer.

This repository also contains a sample integration app which demonstrates how you can integrate the BlinkID into your web app.

BlinkID requires a browser with a support for WebAssembly, but works best with latest versions of Firefox, Chrome, Safari and Microsoft Edge.

Adding BlinkID as a dependency of your project.

The easiest way to add BlinkID as a dev-dependency to your project is by using NPM:

npm install @microblink/blinkid-in-browser-sdk --save-dev

or, if you wish to add a local package instead:

cd /path/to/SDK/folder
npm install
cd /path/to/your/app/folder
npm install /path/to/SDK/folder --save-dev

After adding the @microblink/blinkid-in-browser-sdk as your dev-dependency, make sure to include all files from its build folder in your distribution. Those files contain compiled WebAssembly module and support JS code for loading it, as well as resources needed for BlinkID to work.

The example in the demo app shows how a rollup copy plugin can be used to achieve that. Check the rollup.config.js from the demo app.

Performing your first scan

1. First you will need to create an account at Microblink dashboard where you can generate a demo license for your web app. License is bound to fully qualified domain name of your web app, so please make sure you enter the correct name when asked.

2. Initialize the SDK using following code snippet:

   import * as MicroblinkSDK from '@microblink/blinkid-in-browser-sdk'
   
   // check if browser is supported
   if ( MicroblinkSDK.isBrowserSupported() ) {
       const loadSettings = new MicroblinkSDK.WasmSDKLoadSettings('your-base64-license-key');
   
       MicroblinkSDK.loadWasmModule(loadSettings).then(
           (wasmSDK: MicroblinkSDK.WasmSDK) => {
               // The SDK was initialized successfully, save the wasmSDK
               // for future use
           },
           (error: any) => {
               // error happened during initialization of the WASM SDK
               console.log("Error initializing the SDK: " + error);
           }
       )
   } else {
       console.log("This browser is not supported by the SDK");
   }
   

3. Create recognizer objects that will perform image recognition, configure them and use them to create a RecognizerRunner object:

   import * as MicroblinkSDK from '@microblink/blinkid-in-browser-sdk'
   
   const recognizer = await MicroblinkSDK.createMrtdRecognizer(wasmSDK);
   const recognizerRunner = await MicroblinkSDK.createRecognizerRunner(wasmSDK, [recognizer], true);
   

4. Obtain reference to your HTML video element and create a VideoRecognizer using it and your instance of RecognizerRunner and use it to process:

   const cameraFeed = <HTMLVideoElement>document.getElementById('myCameraVideoElement');
   try {
       const videoRecognizer = await MicroblinkSDK.VideoRecognizer.createVideoRecognizerFromCameraStream(cameraFeed, recognizerRunner);
       const processResult = await videoRecognizer.recognize();
       // obtain recognition results (see next step)
   } catch(error) {
       if (error.name === 'VideoRecognizerError') {
           const reason = (error as MicroblinkSDK.VideoRecognizerError).reason;
           // reason is of type MicroblinkSDK.NotSupportedReason and contains
           // information why video recognizer could not be used.
           // Usually this happens when user didn't give permission to
           // use the camera or when a hardware or OS error occurs.
       }
   }
   

5. If processResult returned from VideoRecognizer's method recognize is not MicroblinkSDK.RecognizerResultState.Empty, then at least one recognizer given to the RecognizerRunner above contains a recognition result. You can extract the result from each recognizer using its getResult method:

   if (processResult !== MicroblinkSDK.RecognizerResultState.Empty) {
       const recognitionResult = await recognizer.getResult();
       console.log(recognitionResult);
   } else {
       console.log("Recognition was not successful!");
   }
   

6. Finally, release the memory on the WebAssembly heap by calling delete method on both RecognizerRunner and each of your recognizers. Also, release the camera stream by calling releaseVideoFeed on instance of VideoRecognizer:

   videoRecognizer.releaseVideoFeed();
   recognizerRunner.delete();
   recognizer.delete();
   

   Note that after releasing those objects it is not valid to call any methods on them, as they are literally destroyed. This is required to release memory resources on WebAssembly heap which are not automatically released with JavaScript's garbage collector. Also note that results returned from getResult method are placed on JavaScript's heap and will be cleaned by garbage collector, just like any other normal JavaScript object.

For more information about available recognizers and RecognizerRunner, see RecognizerRunner and available recognizers.

Recognizing still images

If you just want to perform recognition of still images and do not need live camera recognition, you can do that as well.

1. To do that, first initialize recognizers and RecognizerRunner just as in the steps 1-3 above.

2. Make sure you have the image set to a HTMLImageElement. If you only have the URL of the image that needs recognizing, You can attach it to the image element with following code snippet:

   const imageElement = <HTMLImageElement>document.getElementById('imageToProcess');
   imageElement.src = URL.createObjectURL(imageURL);
   await imageElement.decode();
   

3. Obtain the CapturedFrame object using function captureFrame and give it to the processImage method of the RecognizerRunner:

   const imageFrame = MicroblinkSDK.captureFrame(imageElement);
   const processResult = await recognizerRunner.processImage(imageFrame);
   

4. Proceed as in steps 5. and 6. above. Note that in there is no VideoRecognizer here that needs freeing its resources, but RecognizerRunner and recognizers must be deleted using the delete method.

Optimal deployment of your web app

When browser loads the .wasm file it needs to compile it to the native code. This is unlike javascript code, which is interpreted and compiled to native code only if needed (JIT, a.k.a. Just-in-time compilation). Therefore, before BlinkID is loaded, the browser must download and compile the provided .wasm file. In order to make this faster, you should configure your web server to serve .wasm files with Content-Type: application/wasm. This will instruct the browser that this is a WebAssembly file, which most modern browsers will utilize to perform streaming compilation, i.e. they will start compiling the WebAssembly code as soon as first bytes arrive from the server, instead of waiting for the entire file to download. For more information about streaming compilation, check this article on MDN.

If your server supports serving compressed files, you should utilize that to minimize the download size of your web app. It's easy to notice that .wasm file is not a small file, but it is very compressible. This is also true for all other files that you need to serve for your web app.

For more information about configuring your web server for using compression and for optimal delivery of your web app that uses BlinkID SDK, you should also check the official Emscripten documentation.

The Recognizer concept, RecognizerRunner and VideoRecognizer

This section will first describe what is a Recognizer and how it should be used to perform recognition of the images, videos and camera stream. Next, we will describe what is a RecognizerRunner and how it can be used to tweak the recognition procedure. Finally, a VideoRecognizer will be described and how it builds on top of RecognizerRunner in order to provide support for recognizing a video or a camera stream.

The Recognizer concept

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

The Recognizer is the object on the WebAssembly heap, which means that it will not be automatically cleaned up by the garbage collector once it's not required anymore. Once you are done with using it, you must call the delete method on it to release the memory on the WebAssembly heap. Failing to do so will result in memory leak on the WebAssembly heap which may result with crash of the browser tab running your web app.

Each Recognizer has a Result object, which contains the data that was extracted from the image. The Result for each specific Recognizer can be obtained by calling its getResult method, which will return a Result object placed on the JS heap, i.e. managed by the garbage collector. Therefore, you don't need to call any delete-like methods on the Result object.

Every Recognizer is a stateful object that can be in two possible states: idle state and working state. While in idle state, you are allowed to call method updateSettings which will update its properties according to given settings object. At any time, you can call its currentSettings method to obtain its currently applied settings object. After you create a RecognizerRunner with array containing your recognizer, the state of the Recognizer will change to working state, in which Recognizer object will be used for processing. While being in working state, it is not possible to call method updateSettings (calling it will crash your web app). If you need to change configuration of your recognizer while its being used, you need to call its currentSettings method to obtain its current configuration, update it as you need it, create a new Recognizer of the same type, call updateSettings on it with your modified configuration and finally replace the original Recognizer within the RecognizerRunner by calling its reconfigureRecognizers method. When written as a pseudocode, this would look like:

import * as MicroblinkSDK from '@microblink/blinkid-in-browser-sdk'

// assume myRecognizerInUse is used by the recognizerRunner
const currentSettings = await myRecognizerInUse.currentSettings();
// modify currentSettings as you need
let newRecognizer = await MicroblinkSDK.createRecognizer(); // use appropriate recognizer creation function
await newRecognizer.updateSettings(currentSettings);
// reconfigure recognizerRunner
await recognizerRunner.reconfigureRecognizers([newRecognizer], true); // use `true` or `false` depending of what you want to achieve (see below for the description)
// newRecognizer is now in use and myRecognizerInUse is no longer in use -
// you can delete it if you don't need it anymore
await myRecognizerInUse.delete()

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

RecognizerRunner

The RecognizerRunner is the object that manages the chain of individual Recognizer objects within the recognition process. It must be created by createRecognizerRunner method of the WasmModuleProxy interface, which is a member of WasmSDK interface which is resolved in a promise returned by the loadWasmModule function you've seen above. The function requires a two parameters: an array of Recognizer objects that will be used for processing and a boolean indicating whether multiple Recognizer objects are allowed to have their Results enter the Valid state.

To explain further the boolean parameter, we first need to understand how RecognizerRunner performs image processing.

When the processImage method is called, it processes the image with the first Recognizer in chain. If the Recognizer's Result object changes its state to Valid, then if the above boolean parameter is false, the recognition chain will be broken and promise returned by the method will be immediately resolved. If the above parameter is true, then the image will also be processed with other Recognizer objects in chain, regardless of the state of their Result objects. If, after processing the image with the first Recognizer in chain, its Result object's state is not changed to Valid, the RecognizerRunner will use the next Recognizer object in chain for processing the image and so on - until the end of the chain (if no results become valid or always if above parameter is true) or until it finds the recognizer that has successfully processed the image and changed its Result's state to Valid (if above parameter is false).

You cannot change the order of the Recognizer objects within the chain - no matter the order in which you give Recognizer objects to RecognizerRunner (either to its creation function createRecognizerRunner or to its reconfigureRecognizers method), they are internally ordered in a way that provides best possible performance and accuracy. Also, in order for BlinkID SDK to be able to order Recognizer objects in recognition chain in the best way possible, it is not allowed to have multiple instances of Recognizer objects of the same type within the chain. Attempting to do so will crash your application.

Performing recognition of video streams using VideoRecognizer

Using RecognizerRunner directly could be difficult in cases when you want to perform recognition of the video or the live camera stream. Additionally, handling camera management from the web browser can be sometimes challenging. In order to make this much easier, we provided a VideoRecognizer class.

To perform live camera recognition using the VideoRecognizer, you will need an already set up RecognizerRunner object and a reference to HTMLVideoElement to which camera stream will be attached.

To perform the recognition, you should simply write:

const cameraFeed = <HTMLVideoElement> document.getElementById('cameraFeed');
try {
    const videoRecognizer = await MicroblinkSDK.VideoRecognizer.createVideoRecognizerFromCameraStream(cameraFeed, recognizerRunner);
    const processResult = await videoRecognizer.recognize();
} catch (error) {
    // handle camera error
}

The recognize method of the VideoRecognizer will start the video capture and recognition loop from the camera and will return a Promise that will be resolved when either processImage of the given RecognizerRunner returns Valid for some frame or the timeout given to recognize method is reached (if no timeout is given, a default one is used).

Recognizing a video file

If, instead of performing recognition of live video stream, you want to perform recognition of pre-recorded video file, you should simply construct VideoRecognizer using a different function, as shown below:

const videoRecognizer = await MicroblinkSDK.createVideoRecognizerFromVideoPath(videoPath, htmlVideoElement, recognizerRunner);
const processResult = await videoRecognizer.recognize();

Custom UX with VideoRecognizer

The procedure for using VideoRecognizer described above is quite simple, but has some limits. For example, you can only perform one shot scan with it. As soon as the promise returned by recognize method resolves, the camera feed is paused and you need to start new recognition. However, if you need to perform multiple recognitions in single camera session, without pausing the camera preview, you can use the startRecognition method, as described in the example below;

videoRecognizer.startRecognition
(
    ( recognitionState: MicroblinkSDK.RecognizerResultState ) =>
    {
        // pause recognition before performing any async operation - this will make sure that
        // recognition will not continue while returning the control flow back from this function
        videoRecognizer.pauseRecognition();

        // obtain recognition results directly from recognizers associated with the RecognizerRunner that
        // is associated with the VideoRecognizer

        if ( shouldContinueScanning )
        {
            // resume recognition
            videoRecognizer.resumeRecognition( true );
        }
        else
        {
            // pause the camera feed
            videoRecognizer.pauseVideoFeed();
            // after this line, the VideoRecognizer is in the same state as if promise returned from recognizer was resolved
        }
        // if videoRecognizer is not paused or terminated, after this line the recognition will continue and recognition state
        // will be retained
    }
);

Handling processing events with MetadataCallbacks

Processing events, also known as Metadata callbacks are purely intended for giving processing feedback on UI or to capture some debug information during development of your web app using BlinkID SDK. Callbacks for all events are bundled into the MetadataCallbacks object. We suggest that you check for more information about available callbacks and events to which you can handle in the source code of the MetadataCallbacks interface.

You can associate your implementation of MetadataCallbacks interface with RecognizerRunner either during creation or by invoking its method setMetadataCallbacks. Please note that both those methods need to pass information about available callbacks to the native code and for efficiency reasons this is done at the time setMetadataCallbacks method is called and not every time when change occurs within the MetadataCallbacks object. This means that if you, for example, set onQuadDetection to MetadataCallbacks after you already called setMetadataCallbacks method, the onQuadDetection will not be registered with the native code and therefore it will not be called.

Similarly, if you, for example, remove the onQuadDetection from MetadataCallbacks object after you already called setMetadataCallbacks method, your app will crash in attempt to invoke non-existing function when our processing code attempts to invoke it. We deliberately do not perform null check here because of two reasons:

• it is inefficient
• having no callback, while still being registered to native code is illegal state of your program and it should therefore crash

Remember, each time you make some changes to MetadataCallbacks object, you need to apply those changes to to your RecognizerRunner by calling its setMetadataCallbacks method.

List of available recognizers

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

Success Frame Grabber Recognizer

The SuccessFrameGrabberRecognizer is a special Recognizer that wraps some other Recognizer and impersonates it while processing the image. However, when the Recognizer being impersonated changes its Result into Valid state, the SuccessFrameGrabberRecognizer captures the image and saves it into its own Result object.

Since SuccessFrameGrabberRecognizer impersonates its slave Recognizer object, it is not possible to have both concrete Recognizer object and SuccessFrameGrabberRecognizer that wraps it in the same RecognizerRunner at the same time. Doing so will have the same effect as having multiple instances of the same Recognizer in the same RecognizerRunner - it will crash your application. For more information, see paragraph about RecognizerRunner.

This recognizer is best for use cases when you need to capture the exact image that was being processed by some other Recognizer object at the time its Result became Valid. When that happens, SuccessFrameGrabber's Result will also become Valid and will contain described image. That image will be available in its successFrame property.

ID barcode recognizer

The IdBarcodeRecognizer is recognizer specialized for scanning barcodes from various ID cards.

Machine Readable Travel Document recognizer

The MrtdRecognizer is used for scanning and data extraction from the Machine Readable Zone (MRZ) of the various Machine Readable Travel Documents (MRTDs) like ID cards and passports. This recognizer is not bound to the specific country.

Troubleshooting

Integration problems

In case of problems with the integration of the SDK, first make sure that you have tried integrating the SDK exactly as described in integration instructions. If you have followed the instructions to the letter and you still have the problems, please contact us at help.microblink.com

SDK problems

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

Licensing problems

If you are getting "invalid license key" error or having other license-related problems (e.g. some feature is not enabled that should be), first check the ADB logcat. All license-related problems are logged to web console so it is easy to determine what went wrong.

When you have to determine what is the license-related 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 fully qualified domain name of your app (i.e. where the app is hosted)
• license that is causing problems
• please stress out that you are reporting problem related to WebAssembly version of the BlinkID SDK
• if unsure about the problem, you should also provide excerpt from web console containing the license error

Other problems

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

• Contact us at help.microblink.com describing your problem and provide following information:
  • log from the web console
  • high resolution scan/photo of the item that you are trying to scan
  • information about device and browser that you are using - we need exact version of the browser and operating system it runs on. Also, if it runs on mobile device, we also need the model of the device in question (camera management is specific to both browser, OS and device).
  • please stress out that you are reporting problem related to WebAssembly version of the BlinkID SDK

FAQ and known issues

After switching from trial to production license I get error This entity is not allowed by currently active license! when I create a specific Recognizer object.

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

Additional info

Complete source code of the TypeScript wrapper can be found in here.

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

Changelog

1.0.0-beta.0

• initial beta release of the BlinkID In-browser SDK
• supported recognizers:
  • ID Barcode recognizer for scanning barcodes from various ID documents
  • MRTD recognizer for scanning Machine Readable Zone (MRZ) from Machine Readable Travel Documents (MRTDs), such as IDs and passports