image.binarization (version 0.1.3)

image_binarization: Binarize Images For Enhancing Optical Character Recognition

Description

Binarize images in order to further process it for Optical Character Recognition (OCR) or Handwritten Text Recognition (HTR) purposes

• Otsu - "A threshold selection method from gray-level histograms", 1979.

• Bernsen - "Dynamic thresholding of gray-level images", 1986.

• Niblack - "An Introduction to Digital Image Processing", 1986.

• Sauvola - "Adaptive document image binarization", 1999.

• Wolf - "Extraction and Recognition of Artificial Text in Multimedia Documents", 2003.

• Gatos - "Adaptive degraded document image binarization", 2005. (Partial)

• NICK - "Comparison of Niblack inspired Binarization methods for ancient documents", 2009.

• Su - "Binarization of Historical Document Images Using the Local Maximum and Minimum", 2010.

• T.R. Singh - "A New local Adaptive Thresholding Technique in Binarization", 2011.

• Bataineh - "An adaptive local binarization method for document images based on a novel thresholding method and dynamic windows", 2011. (unreproducible)

• ISauvola - "ISauvola: Improved Sauvola’s Algorithm for Document Image Binarization", 2016.

• WAN - "Binarization of Document Image Using Optimum Threshold Modification", 2018.

Usage

image_binarization(x, type, opts = list())

Value

a binarized image of class magick-image as handled by the magick R package

Arguments

x

an image of class 'magick-image'. In grayscale. E.g. a PGM file. If not provided in grayscale, will extract the gray channel.

type

a character string with the type of binarization to use. Either 'otsu', 'bernsen', 'niblack', 'sauvola', 'wolf', 'nick', 'gatos', 'su', 'trsingh', 'bataineh', 'wan' or 'isauvola'

opts

a list of options to pass on to the algorithm. See the details and the examples.

Details

Options which can be bassed on to the binarization routines, with the defaults between brackets

• otsu: none

• bernsen: window(75L), k(0.2), threshold(100L), contrast-limit(25L)

• niblack: window(75L), k(0.2)

• sauvola: window(75L), k(0.2)

• wolf: window(75L), k(0.2)

• nick: window(75L), k(-0.2)

• gatos: window(75L), k(0.2), glyph(60L)

• su: window(75L), minN(75L)

• trsingh: window(75L), k(0.2)

• bataineh: none

• wan: window(75L), k(0.2)

• isauvola: window(75L), k(0.2)

Note that it is important that you provide the window / threshold / contrast-limit, minN, glyph argument as integers (e.g. as in 75L) and the other parameters as numerics.

Examples

library(magick)
f   <- system.file("extdata", "doxa-example.png", package = "image.binarization")
img <- image_read(f)
img <- image_convert(img, format = "PGM", colorspace = "Gray")

binary <- image_binarization(img, type = "otsu")
binary
binary <- image_binarization(img, type = "bernsen", 
                             opts = list(window = 50L, k = 0.2, threshold = 50L))
binary
binary <- image_binarization(img, type = "niblack", opts = list(window = 75L, k = 0.2))
binary
binary <- image_binarization(img, type = "sauvola")
binary
binary <- image_binarization(img, type = "wolf")
binary
binary <- image_binarization(img, type = "nick", opts = list(window = 75L, k = -0.2))
binary
binary <- image_binarization(img, type = "gatos", opts = list(window = 75L, k = 0.2, glyph = 50L))
binary
binary <- image_binarization(img, type = "su", opts = list(window = 20L))
binary
binary <- image_binarization(img, type = "trsingh")
binary
binary <- image_binarization(img, type = "bataineh")
binary
binary <- image_binarization(img, type = "wan")
binary
binary <- image_binarization(img, type = "isauvola", opts = list(window = 75L, k = 0.2))
binary