Class DdddOcr

Index

Constructors

Methods

classification detection enableDebug setLogSeverityLevel setOcrMode setPath setRanges slideComparison slideMatch

Constructors

constructor

new DdddOcr(): DdddOcr
Class representing an OCR (Optical Character Recognition) model.

Returns DdddOcr
- Defined in src/index.ts:34

Methods

classification

classification(url: string): Promise<string>
Classifies an image by running it through an OCR model.
Parameters
- url: string
Returns Promise<string>
- Defined in src/index.ts:121

detection

detection(url: string): Promise<number[][]>
Detects objects in an image by extracting bounding boxes and optionally visualizing the detection results on the image.

This method reads the image, retrieves bounding boxes, and then optionally draws the detected bounding boxes on the image if debugging is enabled.
Parameters
- url: string
Returns Promise<number[][]>
- Defined in src/index.ts:136

enableDebug

enableDebug(): this
Enables the debug mode and prepares the debug folder.

Returns this
- Defined in src/index.ts:46

setLogSeverityLevel

setLogSeverityLevel(logSeverityLevel: LogSeverityLevel): DdddOcr
Sets the log severity level.
Parameters
- logSeverityLevel: LogSeverityLevel
  Log severity level. See https://github.com/microsoft/onnxruntime/blob/main/include/onnxruntime/core/common/logging/severity.h
Returns DdddOcr
- Defined in src/index.ts:78

setOcrMode

setOcrMode(mode: MODEL_TYPE): this
Sets the OCR mode for the instance.
Parameters
- mode: MODEL_TYPE
Returns this
- Defined in src/index.ts:89

setPath

setPath(root: string): DdddOcr
Sets the root path for ONNX models.
Parameters
- root: string
Returns DdddOcr
- Defined in src/index.ts:65

setRanges

setRanges(charsetRange: string | CHARSET_RANGE): this
Sets the range restriction for OCR results.

This method restricts the characters returned by OCR based on the input:
- For number input, it applies a predefined character set. See the CHARSET_RANGE type for the available options.
- For string input, each character in the string is treated as a valid OCR result.
Parameters
- charsetRange: string | CHARSET_RANGE
Returns this
- Defined in src/index.ts:111

slideComparison

slideComparison(
targetImage: ImageInput,
backgroundImage: ImageInput,
): Promise<SlideComparisonResult>
Finds the slider gap position by comparing two images: one with a gap shadow/highlight visible, and the full background image. Returns the center coordinates of the largest connected difference region.
Parameters
- targetImage: ImageInput
- backgroundImage: ImageInput
Returns Promise<SlideComparisonResult>
- Defined in src/index.ts:162

slideMatch

slideMatch(
    targetImage: ImageInput,
    backgroundImage: ImageInput,
    simpleTarget?: boolean,
): Promise<SlideMatchResult>
Matches a slider puzzle piece against a background image.
- When simpleTarget is false (default), Canny edge detection is applied before template matching — suited for puzzle pieces with transparent backgrounds.
- When simpleTarget is true, grayscale template matching is used directly — suited for simple (opaque) puzzle pieces.
Parameters
- targetImage: ImageInput
- backgroundImage: ImageInput
- simpleTarget: boolean = false
Returns Promise<SlideMatchResult>
- Defined in src/index.ts:149