Documentation
Table of contents

Label Recognizer for Your Website - User Guide

With Dynamsoft Label Recognizer JavaScript edition (DLR-JS), you can add the capability of reading text labels such as passport MRZs, ID cards, VIN numbers, etc. in your web application with just a few lines of code.

version downloads jsdelivr vulnerabilities

Once integrated, your users will be able to launch your website in a web browser, activate their camera functionality, and directly recognize the designated text from the video feed.

In this guide, you will learn step by step on how to integrate this SDK into your website.

Table of Contents

Example Usage

Let’s start by testing an example that demonstrates how to enable a web page to recognize the text from a live video stream.

Check the code

The complete code of the example is shown below

<!DOCTYPE html>
<html>
<body>
  <script src="https://cdn.jsdelivr.net/npm/dynamsoft-label-recognizer-bundle@3.2.3000/dist/dlr.bundle.js"></script>
  <div id="cameraViewContainer" style="width: 100%; height: 60vh"></div>
  <textarea id="results" style="width: 100%; min-height: 10vh; font-size: 3vmin; overflow: auto" disabled></textarea>
  <script>
    Dynamsoft.License.LicenseManager.initLicense("DLS2eyJvcmdhbml6YXRpb25JRCI6IjIwMDAwMSJ9");
    Dynamsoft.Core.CoreModule.loadWasm(["dlr"]);
    (async () => {
      let cvRouter = await Dynamsoft.CVR.CaptureVisionRouter.createInstance();

      let view = await Dynamsoft.DCE.CameraView.createInstance();
      let cameraEnhancer = await Dynamsoft.DCE.CameraEnhancer.createInstance(view);
      document.querySelector("#cameraViewContainer").append(view.getUIElement());
      cvRouter.setInput(cameraEnhancer);

      const resultsContainer = document.querySelector("#results");
      cvRouter.addResultReceiver({ onRecognizedTextLinesReceived: (result) => {
        if (result.textLineResultItems.length > 0) {
          resultsContainer.textContent = "";
          for (let item of result.textLineResultItems) {
            resultsContainer.textContent += `${item.text}\n`;
          }
        }
      }});
      
      let filter = new Dynamsoft.Utility.MultiFrameResultCrossFilter();
      filter.enableResultCrossVerification("text_line", true);
      filter.enableResultDeduplication("text_line", true);
      await cvRouter.addResultFilter(filter);

      await cameraEnhancer.open();
      await cvRouter.startCapturing("RecognizeTextLines_Default");
    })();
  </script>
</body>
</html>

Run via JSFiddle


About the code

  • Dynamsoft.License.LicenseManager.initLicense(): This method initializes the license for using the SDK in the application. Note that the string “DLS2eyJvcmdhbml6YXRpb25JRCI6IjIwMDAwMSJ9” used in this example points to an online license that requires a network connection to work. Read more on Specify the license.

  • Dynamsoft.Core.CoreModule.loadWasm(["dlr"]): This is an optional code. Used to load wasm resources in advance, reducing latency between video playing and label recognizing.

  • Dynamsoft.CVR.CaptureVisionRouter.createInstance(): This method creates a CaptureVisionRouter object cvRouter which controls the entire process in three steps:

    • Retrieve Images from the Image Source
      • cvRouter connects to the image source through the ImageSourceAdapter interface with the method setInput().
        cvRouter.setInput(cameraEnhancer);
        

        The image source in our case is a CameraEnhancer object created with Dynamsoft.DCE.CameraEnhancer.createInstance(view)

    • Coordinate Image-Processing Tasks
      • The coordination happens behind the scenes. cvRouter starts the process by specifying a template “RecognizeTextLines_Default” in the method startCapturing().
        await cvRouter.startCapturing("RecognizeTextLines_Default");
        
    • Dispatch Results to Listening Objects
      • The processing results are returned through the CapturedResultReceiver interface. The CapturedResultReceiver object resultReceiver is registered to cvRouter via the method addResultReceiver().
        cvRouter.addResultReceiver(resultReceiver);
        
      • Also note that reading from video is extremely fast and there could be many duplicate results. We can use a filter with result verification and deduplication enabled to filter out unnecessary results. The object is registered to cvRouter via the method addResultFilter().
        await cvRouter.addResultFilter(filter);
        

Read more on Capture Vision Router.

Run the example

The easiest way to run the example is to use the JSFiddle code editor. You will be asked to allow access to your camera, after which the video will be displayed on the page. After that, you can point the camera at a passport biographical page to read the text in the machine-readable zone.

When the text is recognized, you will see it show up below the video and the location will be highlighted in the video feed.

Alternatively, you can test locally by copying and pasting the code shown above into a local file (e.g. “read-text.html”) and opening it in your browser.

Note:

If you open the web page as file:/// or http:// , the camera may not work correctly because the MediaDevices: getUserMedia() method only works in secure contexts (HTTPS), in some or all supporting browsers.

To make sure your web application can access the camera, please configure your web server to support HTTPS. The following links may help.

  1. NGINX: Configuring HTTPS servers
  2. IIS: How to create a Self Signed Certificate in IIS
  3. Tomcat: Setting Up SSL on Tomcat in 5 minutes
  4. Node.js: npm tls

If the test doesn’t go as expected, you can contact us.

Building your own page

Include the SDK

To utilize the SDK, the initial step involves including the corresponding resource files:

  • core.js encompasses common classes, interfaces, and enumerations that are shared across all Dynamsoft SDKs.
  • license.js introduces the LicenseManager class, which manages the licensing for all Dynamsoft SDKs.
  • utility.js encompasses auxiliary classes that are shared among all Dynamsoft SDKs.
  • dlr.js defines interfaces and enumerations specifically tailored to the label recognizer module.
  • cvr.js introduces the CaptureVisionRouter class, which governs the entire image processing workflow.
  • dce.js comprises classes that offer camera support and basic user interface functionalities.

For simplification, starting from version 3.2.30, we introduced dlr.bundle.js. Including this file is equivalent to incorporating all six packages.

  • dynamsoft-core@3.2.30/dist/core.js
  • dynamsoft-license@3.2.21/dist/license.js
  • dynamsoft-utility@1.2.20/dist/utility.js
  • dynamsoft-label-recognizer@3.2.30/dist/dlr.js
  • dynamsoft-capture-vision-router@2.2.30/dist/cvr.js
  • dynamsoft-camera-enhancer@4.0.2/dist/dce.js

Equivalent to

  • dynamsoft-label-recognizer-bundle@3.2.3000/dist/dlr.bundle.js

In the following chapters, we will use dlr.bundle.js.

Use a public CDN

The simplest way to include the SDK is to use either the jsDelivr or UNPKG CDN. The “hello world” example above uses jsDelivr.

  • jsDelivr

    <script src="https://cdn.jsdelivr.net/npm/dynamsoft-label-recognizer-bundle@3.2.3000/dist/dlr.bundle.js"></script>
    
  • UNPKG

    <script src="https://unpkg.com/dynamsoft-label-recognizer-bundle@3.2.3000/dist/dlr.bundle.js"></script>
    
  • In some rare cases (such as some restricted areas), you might not be able to access the CDN. If this happens, you can use the following files for the test.

    <script src="download2.dynamsoft.com/packages/dynamsoft-label-recognizer-bundle@3.2.3000/dist/dlr.bundle.js"></script>
    

However, please DO NOT use download2.dynamsoft.com resources in a production application as they are for temporary testing purposes only. Instead, you can try hosting the SDK yourself.

Host the SDK yourself

Besides using the public CDN, you can also download the SDK and host its files on your own server or a commercial CDN before including it in your application.

Options to download the SDK:

  • From the website

    Download Dynamsoft Label Recognizer JavaScript Package

  • npm

    npm i dynamsoft-label-recognizer-bundle@3.2.3000 -E
    npm i dynamsoft-capture-vision-std@1.2.10 -E
    npm i dynamsoft-image-processing@2.2.30 -E
    npm i dynamsoft-capture-vision-dnn@1.0.20 -E
    npm i dynamsoft-label-recognizer-data@1.0.11 -E
    
  • yarn

    yarn add dynamsoft-label-recognizer-bundle@3.2.3000 -E
    yarn add dynamsoft-capture-vision-std@1.2.10 -E
    yarn add dynamsoft-image-processing@2.2.30 -E
    yarn add dynamsoft-capture-vision-dnn@1.0.20 -E
    yarn add dynamsoft-label-recognizer-data@1.0.11 -E
    

Depending on how you downloaded the SDK and how you intend to use it, you can typically include it like this:

  • From the website

    <script src="./dynamsoft/distributables/dynamsoft-label-recognizer-bundle@3.2.3000/dist/dlr.bundle.js"></script>
    
  • yarn or npm

    <script src="/node_modules/dynamsoft-label-recognizer-bundle@3.2.3000/dist/dlr.bundle.js"></script>
    

Note:

  • Certain legacy web application servers may lack support for the application/wasm mimetype for .wasm files. To address this, you have two options:
    1. Upgrade your web application server to one that supports the application/wasm mimetype.
    2. Manually define the mimetype on your server. You can refer to the following resources for guidance:
      1. Apache
      2. IIS
      3. Nginx
  • To work properly, the SDK requires a few engine files, which are relatively large and may take quite a few seconds to download. We recommend that you set a longer cache time for these engine files, to maximize the performance of your web application.

    Cache-Control: max-age=31536000
    

    Reference: Cache-Control.

Prepare the SDK

Before using the SDK, you need to configure a few things.

Specify the license

To enable the SDK’s functionality, you must provide a valid license. Utilize the method initLicense() to set your license key.

Dynamsoft.License.LicenseManager.initLicense("DLS2eyJvcmdhbml6YXRpb25JRCI6IjIwMDAwMSJ9");

As previously stated, the key “DLS2eyJvcmdhbml6YXRpb25JRCI6IjIwMDAwMSJ9” serves as a test license valid for 24 hours, applicable to any newly authorized browser. To test the SDK further, you can request a 30-day trial license via the Request a Trial License link.

Upon registering a Dynamsoft account and obtaining the SDK package from the official website, Dynamsoft will automatically create a 30-day free trial license and embed the corresponding license key into all the provided SDK samples.

Specify the location of the “engine” files (optional)

This is usually only required with frameworks like Angular or React, etc. where the referenced JavaScript files such as cvr.js, dlr.js are compiled into another file.

The purpose is to tell the SDK where to find the engine files (*.worker.js, *.wasm.js and *.wasm, etc.). The API is called Dynamsoft.Core.CoreModule.engineResourcePaths:

Object.assign(Dynamsoft.Core.CoreModule.engineResourcePaths, {
  // The following code uses the jsDelivr CDN, feel free to change it to your own location of these files
  core: "https://cdn.jsdelivr.net/npm/dynamsoft-core@3.2.30/dist/",
  license: "https://cdn.jsdelivr.net/npm/dynamsoft-license@3.2.21/dist/",
  dlr: "https://cdn.jsdelivr.net/npm/dynamsoft-label-recognizer@3.2.30/dist/",
  cvr: "https://cdn.jsdelivr.net/npm/dynamsoft-capture-vision-router@2.2.30/dist/",
  dce: "https://cdn.jsdelivr.net/npm/dynamsoft-camera-enhancer@4.0.2/dist/",
  std: "https://cdn.jsdelivr.net/npm/dynamsoft-capture-vision-std@1.2.10/dist/",
  dip: "https://cdn.jsdelivr.net/npm/dynamsoft-image-processing@2.2.30/dist/",
  dnn: "https://cdn.jsdelivr.net/npm/dynamsoft-capture-vision-dnn@1.0.20/dist/",
  // "dlrData" refers to the location of the Convolutional Neural Network (CNN) inference model used for dlr recognition.
  dlrData: "https://cdn.jsdelivr.net/npm/dynamsoft-label-recognizer-data@1.0.11/dist/"
});

Set up and start image processing

Preload the module

The image processing logic is encapsulated within .wasm library files, and these files may require some time for downloading. To enhance the speed, we advise utilizing the following method to preload the libraries.

// Preload the .wasm files
Dynamsoft.Core.CoreModule.loadWasm(["dlr"]);

Create a CaptureVisionRouter object

To use the SDK, we first create a CaptureVisionRouter object.

Dynamsoft.License.LicenseManager.initLicense("DLS2eyJvcmdhbml6YXRpb25JRCI6IjIwMDAwMSJ9");

let cvRouter = null;
try {
    cvRouter = await Dynamsoft.CVR.CaptureVisionRouter.createInstance();
} catch (ex) {
    console.error(ex);
}

Connect an image source

The CaptureVisionRouter object, denoted as cvRouter, is responsible for handling images provided by an image source. In our scenario, we aim to detect text directly from a live video stream. To facilitate this, we initialize a CameraEnhancer object, identified as cameraEnhancer, which is specifically designed to capture image frames from the video feed and subsequently forward them to cvRouter.

To enable video streaming on the webpage, we create a CameraView object referred to as view, which is then passed to cameraEnhancer, and its content is displayed on the webpage.

<div id="cameraViewContainer" style="width: 100%; height: 100vh"></div>
let view = await Dynamsoft.DCE.CameraView.createInstance();
let cameraEnhancer = await Dynamsoft.DCE.CameraEnhancer.createInstance(view);
document.querySelector("#cameraViewContainer").append(view.getUIElement());
cvRouter.setInput(cameraEnhancer);

Register a result receiver

Once the image processing is complete, the results are sent to all the registered CapturedResultReceiver objects. Each CapturedResultReceiver object may encompass one or multiple callback functions associated with various result types. In our particular case, our focus is on recognized text within the images, so it’s sufficient to define the callback function onRecognizedTextLinesReceived:

const resultsContainer = document.querySelector("#results");
const resultReceiver = new Dynamsoft.CVR.CapturedResultReceiver(); 
resultReceiver.onRecognizedTextLinesReceived = (result) => {
  if (result.textLineResultItems.length > 0) {
    resultsContainer.textContent = "";
    for (let item of result.textLineResultItems) {
      resultsContainer.textContent += `${item.text}\n`;
    }
  }
};
cvRouter.addResultReceiver(resultReceiver);

You can also write code like this. It is the same.

cvRouter.addResultReceiver({
  onRecognizedTextLinesReceived: (result) => {
    if (result.textLineResultItems.length > 0) {
      resultsContainer.textContent = "";
      for (let item of result.textLineResultItems) {
        resultsContainer.textContent += `${item.text}\n`;
      }
    }
  },
});

Check out CapturedResultReceiver for more information.

Start the process

With the setup now complete, we can proceed to process the images in two straightforward steps:

  1. Initiate the image source to commence image acquisition. In our scenario, we invoke the open() method on cameraEnhancer to initiate video streaming and simultaneously initiate the collection of images. These collected images will be dispatched to cvRouter as per its request.
  2. Specify a template to commence image processing. In our case, we utilize the RecognizeTextLines_Default template.
await cameraEnhancer.open();
await cvRouter.startCapturing("RecognizeTextLines_Default");

Note:

  • cvRouter is engineered to consistently request images from the image source.
  • The following preset templates are at your disposal.
Template Name Function Description
RecognizeTextLines_Default Identifies and reads any text present.
RecognizeNumbers Specializes in recognizing numerical data.
RecognizeLetters Identifies both uppercase and lowercase English alphabets.
RecognizeNumbersAndLetters Reads both numbers and English alphabets (any case).
RecognizeNumbersAndUppercaseLetters Scans numbers and uppercase English alphabets.
RecognizeUppercaseLetters Focuses on recognizing uppercase English alphabets.

Read more on the preset CaptureVisionTemplates.

Customize the process

Adjust the preset template settings

When making adjustments to some basic tasks, we often only need to modify SimplifiedCaptureVisionSettings.

Retrieve the original image

We have the option to modify the template to retrieve the original image containing the text. For example:

let settings = await cvRouter.getSimplifiedSettings("RecognizeTextLines_Default");
settings.capturedResultItemTypes |= 
  Dynamsoft.Core.EnumCapturedResultItemType.CRIT_ORIGINAL_IMAGE;
await cvRouter.updateSettings("RecognizeTextLines_Default", settings);
await cvRouter.startCapturing("RecognizeTextLines_Default");

Please be aware that it is necessary to update the CapturedResultReceiver object to obtain the original image. For instance:

const EnumCRIT = Dynamsoft.Core.EnumCapturedResultItemType;
resultReceiver.onCapturedResultReceived = (result) => {
  let textLines = result.items.filter(item => item.type === EnumCRIT.CRIT_TEXT_LINE);
  if (textLines.length > 0) {
    let image = result.items.filter(
      item => item.type === EnumCRIT.CRIT_ORIGINAL_IMAGE
    )[0].imageData;
    // The image that we found the text on.
  }
};
Change reading frequency to save power

The SDK is initially configured to process images sequentially without any breaks. Although this setup maximizes performance, it can lead to elevated power consumption, potentially causing the device to overheat. In many cases, it’s possible to reduce the reading speed while still satisfying business requirements. The following code snippet illustrates how to adjust the SDK to process an image every 500 milliseconds:

Please bear in mind that in the following code, if an image’s processing time is shorter than 500 milliseconds, the SDK will wait for the full 500 milliseconds before proceeding to process the next image. Conversely, if an image’s processing time exceeds 500 milliseconds, the subsequent image will be processed immediately upon completion.

let settings = await cvRouter.getSimplifiedSettings("RecognizeTextLines_Default");
settings.minImageCaptureInterval = 500;
await cvRouter.updateSettings("RecognizeTextLines_Default", settings);
await cvRouter.startCapturing("RecognizeTextLines_Default");
Specify a scan region

You can use the parameter roi (region of interest) together with the parameter roiMeasuredInPercentage to configure the SDK to only read a specific region on the image frames. For example:

let settings = await cvRouter.getSimplifiedSettings("RecognizeTextLines_Default");
settings.roiMeasuredInPercentage = true;
settings.roi.points = [
  { x: 5, y: 70 },
  { x: 95, y: 70 },
  { x: 95, y: 90 },
  { x: 5, y: 90 },
];
await cvRouter.updateSettings("RecognizeTextLines_Default", settings);
await cvRouter.startCapturing("RecognizeTextLines_Default");

While the code above accomplishes the task, a more effective approach is to restrict the scan region directly at the image source, as demonstrated in the following code snippet.

  • With the region configured at the image source, the images are cropped right before they are gathered for processing, eliminating the necessity to modify the processing settings further.
  • cameraEnhancer elevates interactivity by overlaying a mask on the video, providing a clear delineation of the scanning region.
  • See also: CameraEnhancer::setScanRegion
cameraEnhancer = await Dynamsoft.DCE.CameraEnhancer.createInstance(view);
cameraEnhancer.setScanRegion({
  x: 5,
  y: 70,
  width: 90,
  height: 20,
  isMeasuredInPercentage: true,
});

Filter the results (Important)

While processing video frames, it’s common for the same text line to be detected multiple times. To enhance the user experience, we can use MultiFrameResultCrossFilter object, having two options currently at our disposal:

Option 1: Verify results across multiple frames
let filter = new Dynamsoft.Utility.MultiFrameResultCrossFilter();
filter.enableResultCrossVerification("text_line", true);
await cvRouter.addResultFilter(filter);

Note:

  • enableResultCrossVerification was designed to cross-validate the outcomes across various frames in a video streaming scenario, enhancing the reliability of the final results.
Option 2: Eliminate redundant results detected within a short time frame
let filter = new Dynamsoft.Utility.MultiFrameResultCrossFilter();
filter.enableResultDeduplication("text_line", true);
await cvRouter.addResultFilter(filter);

Note:

  • enableResultDeduplication was designed to prevent high usage in video streaming scenarios, addressing the repetitive processing of the same text line within a short period of time.

Initially, the filter is set to forget a result 3 seconds after it is first received. During this time frame, if an identical result appears, it is ignored.

Under certain circumstances, this duration can be extended with the method setDuplicateForgetTime().

let filter = new Dynamsoft.Utility.MultiFrameResultCrossFilter();
filter.setDuplicateForgetTime(5000); // Extend the duration to 5 seconds.
await cvRouter.addResultFilter(filter);

You can also enable both options at the same time:

let filter = new Dynamsoft.Utility.MultiFrameResultCrossFilter();
filter.enableResultCrossVerification("text_line", true);
filter.enableResultDeduplication("text_line", true);
filter.setDuplicateForgetTime(5000);
await cvRouter.addResultFilter(filter);

Add feedback

When a text line is detected within the video stream, its position is immediately displayed within the video. Furthermore, utilizing the “Dynamsoft Camera Enhancer” SDK, we can introduce feedback mechanisms, such as emitting a “beep” sound or triggering a “vibration”.

The following code snippet adds a “beep” sound for when a text line is found:

const resultReceiver = new Dynamsoft.CVR.CapturedResultReceiver();
resultReceiver.onRecognizedTextLinesReceived = (result) => {
  if (result.textLineResultItems.length > 0) {
    Dynamsoft.DCE.Feedback.beep();
  }
};
await cvRouter.addResultReceiver(resultReceiver);

Customize the UI

The UI is part of the auxiliary SDK “Dynamsoft Camera Enhancer”, read more on how to customize the UI.

System Requirements

DLR requires the following features to work:

  • Secure context (HTTPS deployment)

    When deploying your application / website for production, make sure to serve it via a secure HTTPS connection. This is required for two reasons

    • Access to the camera video stream is only granted in a security context. Most browsers impose this restriction.

      Some browsers like Chrome may grant the access for http://127.0.0.1 and http://localhost or even for pages opened directly from the local disk (file:///...). This can be helpful for temporary development and test.

    • Dynamsoft License requires a secure context to work.
  • WebAssembly, Blob, URL/createObjectURL, Web Workers

    The above four features are required for the SDK to work.

  • MediaDevices/getUserMedia

    This API is required for in-browser video streaming.

  • getSettings

    This API inspects the video input which is a MediaStreamTrack object about its constrainable properties.

The following table is a list of supported browsers based on the above requirements:

Browser Name Version
Chrome v78+1
Firefox v62+1
Edge v79+
Safari v14+

1 devices running iOS needs to be on iOS 14.3+ for camera video streaming to work in Chrome, Firefox or other Apps using webviews.

Apart from the browsers, the operating systems may impose some limitations of their own that could restrict the use of the SDK. Browser compatibility ultimately depends on whether the browser on that particular operating system supports the features listed above.

Release Notes

Learn about what are included in each release at https://www.dynamsoft.com/label-recognition/docs/web/programming/javascript/release-notes/index.html.

Next Steps

Now that you have got the SDK integrated, you can choose to move forward in the following directions

  1. Check out the Official Samples
  2. Learn about the APIs of the SDK.

This page is compatible for:

Is this page helpful?

YesYes NoNo

In this article:

latest version

  • Latest version(3.2.30)
  • Version 3.x
    • Version 3.0.30
  • Version 2.x
    • Version 2.2.31
    • Version 2.2.30
    • Version 2.2.11
    • Version 2.2.10
    • Version 2.2.4
    • Version 2.2.2
    • Version 2.2.1
    • Version 2.2.0
  • Documentation Homepage
Change +