Detect and Normalize

API Index

Detect and Normalize a Still Images

Detect and Normalize Video Frame Images

detectQuad

Detects quadrilaterals from source data.

detectQuad: (source: Blob | DSImage | DCEFrame | HTMLImageElement | HTMLCanvasElement | string, makeACopy?:boolean) => Promise<Array<DetectedQuadResult>>;

Parameters

• source : specifies the image to detect. The supported image data type include Blob , DSImage , DCEFrame , HTMLImageElement , HTMLCanvasElement and a string which means a URL.
• makeACopy: whether to make a copy of the source data for the detection. Only useful when the source data is type DSImage or DCEFrame in which case if makeACopy is undefined or false, the source data will be lost once it is processed.

Return value

A promise resolving to an array of DetectedQuadResult object that contains all the quad results found in this image.

Code snippet

let quads1 = await normalizer.detectQuad(blob);
let quads2 = await normalizer.detectQuad(url);
let quads3 = await normalizer.detectQuad(htmlImageElement);
let quads4 = await normalizer.detectQuad(htmlCanvasElement);

See also

• Blob
• DSImage
• DCEFrame
• HTMLImageElement
• HTMLCanvasElement

normalize

Normalizes the source image based on the settings in options.

normalize: (source: Blob | DSImage | DCEFrame | HTMLImageElement | HTMLCanvasElement | string, options?: { quad?: Quadrilateral }) => Promise<NormalizedImageResult>;

Parameters

source : specifies the image data to normalize. The supported image data type include Blob , DSImage , DCEFrame , HTMLImageElement , HTMLCanvasElement and a string which means an URL.

options : specifies how normalization should be done. If “quad” is passed, it means that the image should be cropped according to the quad, and the resulting image should also be “deskewed” and “perspective corrected”. More options to be added for the normalization in the future.

Return value

A promise resolving to a NormalizedImageResult object that contains the normalized result of this image.

Code snippet

let quads = await normalizer.detectQuad(blob);
let normalizedImageResult = await normalizer.normalize(sourceImage, {
    quad: quads[0].location
});

setImageSource

Sets an image source for continous scanning.

setImageSource(imageSource: ImageSource, options?: object): Promise<void>;

Parameters

imageSource : specifies the image source.

options : options to help with the usage of the ImageSource object. At present, it only contains one property resultsHighlightBaseShapes that accepts Dynamsoft.DCE.DrawingItem as its value to help with the highlighting of text regions as shown in the code snippet below. More properties will be added as needed in the future.

Return value

A promise that resolves when the operation succeeds.

Code snippet

let normalizer = await Dynamsoft.DDN.DocumentNormalizer.createInstance();
let enhancer = await Dynamsoft.DCE.CameraEnhancer.createInstance();
let options = {
    resultsHighlightBaseShapes: Dynamsoft.DCE.DrawingItem
};
await normalizer.setImageSource(enhancer, options);
normalizer.onQuadDetected = async (quadResults, sourceImage) => {
    //...
}
await normalizer.startScanning(true);

onQuadDetected

This event is triggered when a new quadrilateral is detected.

onQuadDetected: (quadResults: Array<DetectedQuadResult>, sourceImage: Blob | DSImage | DCEFrame | HTMLImageElement | HTMLCanvasElement | string) => void;

Parameters

quadResults : an array of DetectedQuadResult that hold the detected quad results.

sourceImage : specifies the image to detect. The supported image data type include Blob , DSImage , DCEFrame , HTMLImageElement , HTMLCanvasElement and a string which means an URL.

Code snippet

normalizer.onQuadDetected = async (quadResults, sourceImage) => {
    for (let quadResult of quadResults) {
        let normalizedImageResult = await normalizer.normalize(sourceImage, {
            quad: quadResult.location
        });
    }
}

setQuadResultFilter

Sets a function to filter a detected quadrilateral.

setQuadResultFilter: (filter: (quad: DetectedQuadResult) => boolean) => void; 

Parameters

filter : a function which is used to filter a DetectedQuadResult and return a boolean to indicate whether it’s wanted.

Code snippet

normalizer.setQuadResultFilter((quadResult) => {
    if (quadResult.confidenceAsDocumentBoundary > 70) {
        return true;
    }
});

confirmQuadForNormalization

Pauses the video stream and enter “editor mode” where the quadrilaterals in current frame are selectable and editable. Then you can select one quadrilateral that contains the content you are interested in and edit it to a desirable shape to be normalized.

confirmQuadForNormalization: () => void;

Code snippet

normalizer.confirmQuadForNormalization();

normalizeWithConfirmedQuad

Normalizes the image with the one quadrilateral which is confirmed in confirmQuadForNormalization() .

normalizeWithConfirmedQuad: () => Promise<NormalizedImageResult>;

Return value

A promise resolving to a NormalizedImageResult object that contains the normalized result of this image.

Code snippet

let normalizedImageResult = await normalizer.normalizeWithConfirmedQuad();
if (normalizedImageResult) {
    // Show the normalized image in a Canvas
    const cvs = normalizedImageResult.image.toCanvas();
    document.querySelector("#normalized-result").appendChild(cvs);
}

startScanning

Open the camera and starts continuous scanning of incoming images.

startScanning(appendOrShowUI?: boolean): Promise<void>;

Parameters

appendOrShowUI : this parameter specifies how to handle the UI that comes with the bound CameraEnhancer instance. When set to true, if the UI doesn’t exist in the DOM tree, the CameraEnhancer instance will append it in the DOM and show it; if the UI already exists in the DOM tree but is hidden, it’ll be displayed. When not set or set to false, it means not to change the original state of that UI: if it doesn’t exist in the DOM tree, nothing shows up on the page; if it exists in the DOM tree, it may or may not show up depending on its original state.

Return value

A promise that resolves when the operation succeeds.

Code Snippet

let normalizer = await Dynamsoft.DDN.DocumentNormalizer.createInstance();
let enhancer = await Dynamsoft.DCE.CameraEnhancer.createInstance();
let options = {
    resultsHighlightBaseShapes: Dynamsoft.DCE.DrawingItem
};
await normalizer.setImageSource(enhancer, options);
normalizer.onQuadDetected = async (quadResults, sourceImage) => {
    for (let quadResult of quadResults) {
        let normalizedImageResult = await normalizer.normalize(sourceImage, {
            quad: quadResult.location
        });
    }
}
normalizer.startScanning(true);

pauseScanning

Pause continuous scanning but keep the video stream.

Parameters

options : Options to configure how the pause works. At present, it only contains one property keepResultsHighlighted which, when set to true, will keep the text found on the frame (at the time of the pause) highlighted.

pauseScanning(options?: object): void;

resumeScanning

Resumes continuous scanning.

resumeScanning(): void;

stopScanning

Stops continuous scanning and closes the video stream.

stopScanning(hideUI?: boolean): void;

Parameters

hideUI : this parameter specifies how to handle the UI that comes with the bound CameraEnhancer instance. When set to true, if the UI doesn’t exist in the DOM tree or it exists but is hidden, nothing is done; if the UI already exists in the DOM tree and is shown, it’ll be hidden. When not set or set to false, it means not to change the original state of that UI: if it doesn’t exist in the DOM tree, nothing happens; if it exists in the DOM tree, it may or may not be hidden depending on its original state.

Code Snippet

normalizer.stopScanning(true);