strokeWidthTransform
src/loadSWT.jsGet the bounding boxes over the text from the image using stroke width transform (SWT)
Parameters
image
(Image)
options
(object)
| Name | Description |
|---|---|
options.scaleInvariant boolean
(default true)
|
Enable scale invariant SWT (to scale to different sizes and then combine the results) |
options.interval number
(default 1)
|
Intervals for the scale invariant option |
options.minNeighbors number
(default 1)
|
Minimal neighbors to make a detection valid, this is for scale-invariant version. |
options.size number
(default 3)
|
Size of the sobel operator for Canny Edge. |
options.lowThresh number
(default 124)
|
Low threshold for Canny Edge. |
options.highThresh number
(default 204)
|
High threshold for Canny Edge. |
options.maxHeight number
(default 300)
|
Maximum height for a letter. |
options.minHeight number
(default 8)
|
Minimum height for a letter. |
options.minArea number
(default 38)
|
Minimum occupied area for a letter. |
options.aspectRatio number
(default 8)
|
Maximum aspect ratio for a letter. |
options.stdRatio number
(default 0.83)
|
The inner-class standard derivation when grouping letters. |
options.thicknessRatio number
(default 1.5)
|
The allowable thickness variance when grouping letters. |
options.heightRatio number
(default 1.7)
|
The allowable height variance when grouping letters. |
options.intensityThresh number
(default 31)
|
The allowable intensity variance when grouping letters. |
options.distanceRatio number
(default 2.9)
|
The allowable distance variance when grouping letters. |
options.intersectRatio number
(default 1.3)
|
The allowable intersect variance when grouping letters. |
options.elongateRatio number
(default 1.9)
|
The allowable elongate variance when grouping letters. |
options.letterThresh number
(default 3)
|
The allowable letter threshold. |
options.breakdown boolean
(default true)
|
If breakdown text line into words. |
options.breakdownRatio number
(default 1)
|
Apply OTSU method and if inter-class variance above the threshold, it will be break down into words. |
Returns
Array<Roi>:
Array of regions that contains text.