Mobile Document Scanner API Reference / DocumentNormalizerView

Class: DocumentNormalizerView

Defined in: src/views/DocumentCorrectionView.ts:191

Constructors

Constructor

new DocumentNormalizerView(resources, config, scannerView?): DocumentCorrectionView

Defined in: src/views/DocumentCorrectionView.ts:222

Parameters

resources

SharedResources

config

DocumentCorrectionViewConfig

scannerView?

DocumentScannerView

Returns

DocumentCorrectionView

Methods

confirmCorrection()

confirmCorrection(): Promise<void>

Defined in: src/views/DocumentCorrectionView.ts:839

Confirm the boundary adjustments and apply perspective correction to the image.

Returns

Promise<void>

Throws

If no quadrilateral boundary is found on the drawing layer

Remarks

Retrieves boundary, performs correction via correctImage, updates result, invokes callback, resolves promise.

---

correctImage()

correctImage(points): Promise<DeskewedImageResultItem | undefined>

Defined in: src/views/DocumentCorrectionView.ts:927

Apply perspective correction to the document image using Dynamsoft Document Normalizer (DDN).

Parameters

points

[Point, Point, Point, Point]

The quadrilateral corner points defining the document boundary

Returns

Promise<DeskewedImageResultItem | undefined>

The perspective-corrected (deskewed) image, or undefined if correction fails

Remarks

Configures ROI with quadrilateral points, processes image with normalization template.

---

dispose()

dispose(preserveResolver?): void

Defined in: src/views/DocumentCorrectionView.ts:963

Clean up and release resources.

Parameters

preserveResolver?

boolean = false

Whether to preserve the currentCorrectionResolver promise resolver

Returns

void

Remarks

Disposes ImageEditorView, clears layer and container. Optionally preserves resolver for handleRetake.

---

hideView()

hideView(): void

Defined in: src/views/DocumentCorrectionView.ts:911

Hide the correction view by setting its container display to "none".

Returns

void

Remarks

Sets container display to "none" without disposing resources.

---

initialize()

initialize(): Promise<void>

Defined in: src/views/DocumentCorrectionView.ts:233

Returns

Promise<void>

---

launch()

launch(): Promise<DocumentResult>

Defined in: src/views/DocumentCorrectionView.ts:870

Returns

Promise<DocumentResult>

---

setBoundaryAutomatically()

setBoundaryAutomatically(): Promise<void>

Defined in: src/views/DocumentCorrectionView.ts:796

Automatically detect document boundaries using Dynamsoft Document Normalizer (DDN).

Returns

Promise<void>

Remarks

This method re-runs document boundary detection on the current image using the CaptureVisionRouter and DDN (Dynamsoft Document Normalizer) engine. The process:

1. Initializes settings from DocumentCorrectionViewConfig.templateFilePath if provided
2. Retrieves the detection template settings via DocumentCorrectionViewConfig.utilizedTemplateNames
3. Configures the router to:
   • Output the original image for further processing
   • Process images at full resolution (no downscaling via maxImageSideLength = Infinity)
4. Captures and analyzes the image to detect document boundaries
5. Handles the detection result:
   • Boundary detected: Creates a QuadDrawingItem from the detected Quadrilateral and adds it to the layer via addQuadToLayer
   • No boundary detected: Falls back to setFullImageBoundary to use full image dimensions

This allows users to retry automatic detection if:

• Manual adjustments were made but proved unsatisfactory
• Initial detection failed due to lighting or positioning issues that were later corrected
• The image was uploaded without initial detection

Triggered by clicking the "Detect Borders" button created in createControls.

See

• CaptureVisionRouter - Processes the image for boundary detection
• DetectedQuadResultItem - Contains the detected document boundaries
• addQuadToLayer - Adds the detected quadrilateral to the drawing layer
• setFullImageBoundary - Fallback when no boundaries are detected
• SharedResources.cvRouter - The router instance used for detection

---

setFullImageBoundary()

setFullImageBoundary(): void

Defined in: src/views/DocumentCorrectionView.ts:743

Reset the document boundary to match the full image dimensions.

Returns

void

Throws

If no captured image is available in SharedResources.result

Remarks

This method creates a quadrilateral boundary that encompasses the entire image, effectively disabling perspective correction. The quadrilateral corners are positioned at:

• Top-left: (0, 0)
• Top-right: (width, 0)
• Bottom-right: (width, height)
• Bottom-left: (0, height)

This is useful when:

• The document occupies the entire image with no background visible
• Automatic boundary detection failed or detected incorrect boundaries
• Users want to skip perspective correction and use the original image dimensions

The created quadrilateral is added to the DrawingLayer via addQuadToLayer, replacing any existing boundary. Users can still manually adjust the corners after reset.

Triggered by clicking the "Full Image" button created in createControls.

See

• SharedResources.result - Contains the original image dimensions
• QuadDrawingItem - The quadrilateral type added to the drawing layer
• addQuadToLayer - Adds the full-image quadrilateral to the layer
• setBoundaryAutomatically - Alternative method for automatic boundary detection