GetConfidenceColors: Get Colors for Displaying Confidence of OCR

In duncantl/Rtesseract: Interface to the tesseract OCR system

Get Colors for Displaying Confidence of OCR

Description

This is a convenience function for computing the color values to indicate the confidence/certainty of each of the recognized elements in the OCR. This computes the color palette, by default, and then maps the confidence values into bins. By default, the more green the value, the more certainty. A caller can specify different end points of the color range (via colorEnds).

Usage

GetConfidenceColors(bbox, confidences = bbox[, "confidence"],
                    numColors = 10,
                    colors = colorRampPalette(colorEnds)(numColors),
                    colorEnds = getOption("OCRConfidenceColors", c("red", "lightgreen")),  
                    intervals = quantile(confidences, seq(0, 1, by = 1/numColors)))

Arguments

bbox	the matrix of matched elements returned from GetBoxes
confidences	the vector of confidence values. These are extracted from bbox or can be specified directly as the return value from GetConfidences
numColors	the number of colors to use in the color palette
colors	a vector of color values (as a character vector) in case the caller wants to specify the colors directly.
colorEnds	a character vector of two colors giving the color for the minimum and maximum confidence. These are used to interpolate the other color values for the confidences. The caller can specify these two color values via the R session option named OCRConfidenceColors.
intervals	the bins by which to group the confidences

Value

A character vector of colors corresponding to the rows in bbox or elements in confidences

Author(s)

Duncan Temple Lang

See Also

GetConfidences, GetBoxes, plot

Examples

f = system.file("trainingSample", "eng.tables.exp0.png", package = "Rtesseract")
ts = tesseract(f)
bbox = GetBoxes(ts)
colors = GetConfidenceColors(bbox) 
plot(ts, bbox = bbox, border = colors)

m = plot(ts, img = NULL, bbox = bbox, border = colors)
rect(m[,1], m[,2], m[,3], m[,4], col = colors)

duncantl/Rtesseract documentation built on March 25, 2022, 5:50 a.m.