Package 'magick'

Title: Advanced Graphics and Image-Processing in R
Description: Bindings to 'ImageMagick': the most comprehensive open-source image processing library available. Supports many common formats (png, jpeg, tiff, pdf, etc) and manipulations (rotate, scale, crop, trim, flip, blur, etc). All operations are vectorized via the Magick++ STL meaning they operate either on a single frame or a series of frames for working with layers, collages, or animation. In RStudio images are automatically previewed when printed to the console, resulting in an interactive editing environment. The latest version of the package includes a native graphics device for creating in-memory graphics or drawing onto images using pixel coordinates.
Authors: Jeroen Ooms [aut, cre]
Maintainer: Jeroen Ooms <[email protected]>
License: MIT + file LICENSE
Version: 2.8.5
Built: 2024-11-08 06:16:39 UTC
Source: https://github.com/ropensci/magick

Help Index


Magick Image Processing

Description

The magick package for graphics and image processing in R. Important resources:

Details

Documentation is split into the following pages:

  • analysis - metrics and calculations: compare, fft

  • animation - manipulate or combine multiple frames: animate, morph, mosaic, montage, average, append, apply

  • attributes - image properties: comment, info

  • color - contrast, brightness, colors: modulate, quantize, map, transparent, background, colorize, contrast, normalize, enhance, equalize, median

  • composite - advanced joining: composite, border, frame

  • device - creating graphics and drawing on images

  • editing - basic image IO: read, write, convert, join, display, brose

  • effects - fun effects: despecle, reducenoise, noise, blur, charcoal, edge, oilpaint, emboss, implode, negate

  • geometry - specify points, areas and sizes using geometry syntax

  • ocr - extract text from image using tesseract package

  • options - list option types and values supported in your version of ImageMagick

  • painting - flood fill and annotating text

  • transform - shape operations: trim, chop, rotate, resize, scale, sample crop, flip, flop, deskew, page

See Also

Other image: analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video


Image Analysis

Description

Functions for image calculations and analysis. This part of the package needs more work.

Usage

image_compare(image, reference_image, metric = "", fuzz = 0)

image_compare_dist(image, reference_image, metric = "", fuzz = 0)

image_fft(image)

Arguments

image

magick image object returned by image_read() or image_graph()

reference_image

another image to compare to

metric

string with a metric from metric_types() such as "AE" or "phash"

fuzz

relative color distance (value between 0 and 100) to be considered similar in the filling algorithm

Details

For details see Image++ documentation. Short descriptions:

  • image_compare calculates a metric by comparing image with a reference image.

  • image_fft returns Discrete Fourier Transform (DFT) of the image as a magnitude / phase image pair. I wish I knew what this means.

Here image_compare() is vectorized over the first argument and returns the diff image with the calculated distortion value as an attribute.

See Also

Other image: _index_, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

out1 <- image_blur(logo, 3)
out2 <- image_oilpaint(logo, 3)
input <- c(logo, out1, out2, logo)
if(magick_config()$version >= "6.8.7"){
  diff_img <- image_compare(input, logo, metric = "AE")
  attributes(diff_img)
}

Image Frames and Animation

Description

Operations to manipulate or combine multiple frames of an image. Details below.

Usage

image_animate(
  image,
  fps = 10,
  delay = NULL,
  loop = 0,
  dispose = c("background", "previous", "none"),
  optimize = FALSE
)

image_coalesce(image)

image_morph(image, frames = 8)

image_mosaic(image, operator = NULL)

image_flatten(image, operator = NULL)

image_average(image)

image_append(image, stack = FALSE)

image_apply(image, FUN, ...)

image_montage(
  image,
  geometry = NULL,
  tile = NULL,
  gravity = "Center",
  bg = "white",
  shadow = FALSE
)

Arguments

image

magick image object returned by image_read() or image_graph()

fps

frames per second. Ignored if delay is not NULL.

delay

delay after each frame, in 1/100 seconds. Must be length 1, or number of frames. If specified, then fps is ignored.

loop

how many times to repeat the animation. Default is infinite.

dispose

a frame disposal method from dispose_types()

optimize

optimize the gif animation by storing only the differences between frames. Input images must be exactly the same size.

frames

number of frames to use in output animation

operator

string with a composite operator from compose_types()

stack

place images top-to-bottom (TRUE) or left-to-right (FALSE)

FUN

a function to be called on each frame in the image

...

additional parameters for FUN

geometry

a geometry string that defines the size the individual thumbnail images, and the spacing between them.

tile

a geometry string for example "4x5 with limits on how the tiled images are to be laid out on the final result.

gravity

a gravity direction, if the image is smaller than the frame, where in the frame is the image to be placed.

bg

a background color string

shadow

enable shadows between images

Details

For details see Magick++ STL documentation. Short descriptions:

  • image_animate coalesces frames by playing the sequence and converting to gif format.

  • image_morph expands number of frames by interpolating intermediate frames to blend into each other when played as an animation.

  • image_mosaic inlays images to form a single coherent picture.

  • image_montage creates a composite image by combining frames.

  • image_flatten merges frames as layers into a single frame using a given operator.

  • image_average averages frames into single frame.

  • image_append stack images left-to-right (default) or top-to-bottom.

  • image_apply applies a function to each frame

The image_apply function calls an image function to each frame and joins results back into a single image. Because most operations are already vectorized this is often not needed. Note that FUN() should return an image. To apply other kinds of functions to image frames simply use lapply, vapply, etc.

See Also

Other image: _index_, analysis, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# Combine images
logo <- image_read("https://jeroen.github.io/images/Rlogo.png")
oldlogo <- image_read("https://jeroen.github.io/images/Rlogo-old.png")

# Create morphing animation
both <- image_scale(c(oldlogo, logo), "400")
image_average(image_crop(both))
image_animate(image_morph(both, 10))

# Create thumbnails from GIF
banana <- image_read("https://jeroen.github.io/images/banana.gif")
length(banana)
image_average(banana)
image_flatten(banana)
image_append(banana)
image_append(banana, stack = TRUE)

# Append images together
wizard <- image_read("wizard:")
image_append(image_scale(c(image_append(banana[c(1,3)], stack = TRUE), wizard)))

image_composite(banana, image_scale(logo, "300"))

# Break down and combine frames
front <- image_scale(banana, "300")
background <- image_background(image_scale(logo, "400"), 'white')
frames <- image_apply(front, function(x){image_composite(background, x, offset = "+70+30")})
image_animate(frames, fps = 10)
# Simple 4x3 montage
input <- rep(logo, 12)
image_montage(input, geometry = 'x100+10+10', tile = '4x3', bg = 'pink', shadow = TRUE)

# With varying frame size
input <- c(wizard, wizard, logo, logo)
image_montage(input, tile = '2x2', bg = 'pink', gravity = 'southwest')

Convert to EBImage

Description

Convert a Magick image to EBImage class. Note that EBImage only supports multi-frame images in greyscale.

Usage

as_EBImage(image)

Arguments

image

magick image object returned by image_read() or image_graph()


Image Attributes

Description

Attributes are properties of the image that might be present on some images and might affect image manipulation methods.

Usage

image_comment(image, comment = NULL)

image_info(image)

Arguments

image

magick image object returned by image_read() or image_graph()

comment

string to set an image comment

Details

Each attribute can be get and set with the same function. The image_info() function returns a data frame with some commonly used attributes.

See Also

Other image: _index_, analysis, animation, color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video


RStudio Graphics AutoViewer

Description

This enables a addTaskCallback that automatically updates the viewer after the state of a magick graphics device has changed. This is enabled by default in RStudio.

Usage

autoviewer_enable()

autoviewer_disable()

Examples

# Only has effect in RStudio (or other GUI with a viewer):
autoviewer_enable()

img <- magick::image_graph()
plot(1)
abline(0, 1, col = "blue", lwd = 2, lty = "solid")
abline(0.1, 1, col = "red", lwd = 3, lty = "dotted")

autoviewer_disable()
abline(0.2, 1, col = "green", lwd = 4, lty = "twodash")
abline(0.3, 1, col = "black", lwd = 5, lty = "dotdash")

autoviewer_enable()
abline(0.4, 1, col = "purple", lwd = 6, lty = "dashed")
abline(0.5, 1, col = "yellow", lwd = 7, lty = "longdash")

Magick Configuration

Description

ImageMagick can be configured to support various additional tool and formats via external libraries. These functions show which features ImageMagick supports on your system.

Usage

coder_info(format)

magick_config()

magick_set_seed(seed)

Arguments

format

image format such as png, tiff or pdf.

seed

integer with seed value to use

Details

Note that coder_info raises an error for unsupported formats.

References

https://www.imagemagick.org/Magick++/CoderInfo.html

Examples

coder_info("png")
coder_info("jpg")
coder_info("pdf")
coder_info("tiff")
coder_info("gif")
# Reproduce random image
magick_set_seed(123)
image_blank(200,200, pseudo_image = "plasma:fractal")

Image Color

Description

Functions to adjust contrast, brightness, colors of the image. Details below.

Usage

image_modulate(image, brightness = 100, saturation = 100, hue = 100)

image_quantize(
  image,
  max = 256,
  colorspace = "rgb",
  dither = TRUE,
  treedepth = NULL
)

image_map(image, map, dither = FALSE)

image_ordered_dither(image, threshold_map)

image_channel(image, channel = "lightness")

image_separate(image, channel = "default")

image_combine(image, colorspace = "sRGB", channel = "default")

image_transparent(image, color, fuzz = 0)

image_background(image, color, flatten = TRUE)

image_colorize(image, opacity, color)

image_contrast(image, sharpen = 1)

image_normalize(image)

image_enhance(image)

image_equalize(image)

image_median(image, radius = 1)

Arguments

image

magick image object returned by image_read() or image_graph()

brightness

modulation of brightness as percentage of the current value (100 for no change)

saturation

modulation of saturation as percentage of the current value (100 for no change)

hue

modulation of hue is an absolute rotation of -180 degrees to +180 degrees from the current position corresponding to an argument range of 0 to 200 (100 for no change)

max

preferred number of colors in the image. The actual number of colors in the image may be less than your request, but never more.

colorspace

string with a colorspace from colorspace_types for example "gray", "rgb" or "cmyk"

dither

a boolean (defaults to TRUE) specifying whether to apply Floyd/Steinberg error diffusion to the image: averages intensities of several neighboring pixels

treedepth

depth of the quantization color classification tree. Values of 0 or 1 allow selection of the optimal tree depth for the color reduction algorithm. Values between 2 and 8 may be used to manually adjust the tree depth.

map

reference image to map colors from

threshold_map

A string giving the dithering pattern to use. See the ImageMagick documentation for possible values

channel

a string with a channel from channel_types for example "alpha" or "hue" or "cyan"

color

a valid color string such as "navyblue" or "#000080". Use "none" for transparency.

fuzz

relative color distance (value between 0 and 100) to be considered similar in the filling algorithm

flatten

should image be flattened before writing? This also replaces transparency with background color.

opacity

percentage of opacity used for coloring

sharpen

enhance intensity differences in image

radius

replace each pixel with the median color in a circular neighborhood

Details

For details see Magick++ STL documentation. Short descriptions:

  • image_modulate adjusts brightness, saturation and hue of image relative to current.

  • image_quantize reduces number of unique colors in the image.

  • image_ordered_dither reduces number of unique colors using a dithering threshold map.

  • image_map replaces colors of image with the closest color from a reference image.

  • image_channel extracts a single channel from an image and returns as grayscale.

  • image_transparent sets pixels approximately matching given color to transparent.

  • image_background sets background color. When image is flattened, transparent pixels get background color.

  • image_colorize overlays a solid color frame using specified opacity.

  • image_contrast enhances intensity differences in image

  • image_normalize increases contrast by normalizing the pixel values to span the full range of colors

  • image_enhance tries to minimize noise

  • image_equalize equalizes using histogram equalization

  • image_median replaces each pixel with the median color in a circular neighborhood

Note that colors are also determined by image properties imagetype and colorspace which can be modified via image_convert().

See Also

Other image: _index_, analysis, animation, attributes(), composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# manually adjust colors
logo <- image_read("logo:")
image_modulate(logo, brightness = 200)
image_modulate(logo, saturation = 150)
image_modulate(logo, hue = 200)

# Reduce image to 10 different colors using various spaces
image_quantize(logo, max = 10, colorspace = 'gray')
image_quantize(logo, max = 10, colorspace = 'rgb')
image_quantize(logo, max = 10, colorspace = 'cmyk')

image_ordered_dither(logo, 'o8x8')
# Change background color
translogo <- image_transparent(logo, 'white')
image_background(translogo, "pink", flatten = TRUE)

# Compare to flood-fill method:
image_fill(logo, "pink", fuzz = 20)

# Other color tweaks
image_colorize(logo, 50, "red")
image_contrast(logo)
image_normalize(logo)
image_enhance(logo)
image_equalize(logo)
image_median(logo)

# Alternate way to convert into black-white
image_convert(logo, type = 'grayscale')

Image Composite

Description

Similar to the ImageMagick composite utility: compose an image on top of another one using a CompositeOperator.

Usage

image_composite(
  image,
  composite_image,
  operator = "atop",
  offset = "+0+0",
  gravity = "northwest",
  compose_args = ""
)

image_border(image, color = "lightgray", geometry = "10x10", operator = "copy")

image_frame(image, color = "lightgray", geometry = "25x25+6+6")

image_shadow_mask(image, geometry = "50x10+30+30")

image_shadow(
  image,
  color = "black",
  bg = "none",
  geometry = "50x10+30+30",
  operator = "copy",
  offset = "+20+20"
)

image_shade(image, azimuth = 30, elevation = 30, color = FALSE)

Arguments

image

magick image object returned by image_read() or image_graph()

composite_image

composition image

operator

string with a composite operator from compose_types()

offset

string with either a gravity_type or a geometry_point to set position of top image.

gravity

string with gravity value from gravity_types.

compose_args

additional arguments needed for some composite operations

color

Set to true to shade the red, green, and blue components of the image.

geometry

a geometry string to set height and width of the border, e.g. "10x8". In addition image_frame allows for adding shadow by setting an offset e.g. "20x10+7+2".

bg

background color

azimuth

position of light source

elevation

position of light source

Details

The image_composite function is vectorized over both image arguments: if the first image has n frames and the second m frames, the output image will contain n * m frames.

The image_border function creates a slightly larger solid color frame and then composes the original frame on top. The image_frame function is similar but has an additional feature to create a shadow effect on the border (which is really ugly).

See Also

Other image: _index_, analysis, animation, attributes(), color, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# Compose images using one of many operators
imlogo <- image_scale(image_read("logo:"), "x275")
rlogo <- image_read("https://jeroen.github.io/images/Rlogo-old.png")

# Standard is atop
image_composite(imlogo, rlogo)

# Same as 'blend 50' in the command line
image_composite(imlogo, rlogo, operator = "blend", compose_args="50")

# Offset can be geometry or gravity
image_composite(logo, rose, offset = "+100+100")
image_composite(logo, rose, gravity = "East")

# Add a border frame around the image
image_border(imlogo, "red", "10x10")
image_frame(imlogo)
image_shadow(imlogo)
image_shade(imlogo)

Set encoder defines

Description

So called 'defines' are properties that are passed along to external filters and libraries. Usually defines are used in image_read or image_write to control the image encoder/decoder, but you can also set these manually on the image object.

Usage

image_set_defines(image, defines)

Arguments

image

magick image object returned by image_read() or image_graph()

defines

a named character vector with extra options to control reading. These are the ⁠-define key{=value}⁠ settings in the command line tool. Use an empty string for value-less defines, and NA to unset a define.

Details

The defines values must be a character string, where the names contain the defines keys. Each name must be of the format "enc:key" where the first part is the encoder or filter to which the key is passed. For example "png:...." defines can control the encoding and decoding of png images.

The image_set_defines function does not make a copy of the image, so the defined values remain in the image object until they are overwritten or unset.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# Write an image
x <- image_read("https://jeroen.github.io/images/frink.png")
image_write(x, "frink.png")

# Pass some properties to PNG encoder
defines <- c("png:compression-filter" = "1", "png:compression-level" = "0")
image_set_defines(x, defines)
image_write(x, "frink-uncompressed.png")

# Unset properties
defines[1:2] = NA
image_set_defines(x, defines)
image_write(x, "frink-final.png")

# Compare size and cleanup
file.info(c("frink.png", "frink-uncompressed.png", "frink-final.png"))
unlink(c("frink.png", "frink-uncompressed.png", "frink-final.png"))

Magick Graphics Device

Description

Graphics device that produces a Magick image. Can either be used like a regular device for making plots, or alternatively via image_draw to open a device which draws onto an existing image using pixel coordinates. The latter is vectorized, i.e. drawing operations are applied to each frame in the image.

Usage

image_graph(
  width = 800,
  height = 600,
  bg = "white",
  pointsize = 12,
  res = 72,
  clip = TRUE,
  antialias = TRUE
)

image_draw(image, pointsize = 12, res = 72, antialias = TRUE, ...)

image_capture()

Arguments

width

in pixels

height

in pixels

bg

background color

pointsize

size of fonts

res

resolution in pixels

clip

enable clipping in the device. Because clipping can slow things down a lot, you can disable it if you don't need it.

antialias

TRUE/FALSE: enables anti-aliasing for text and strokes

image

an existing image on which to start drawing

...

additional device parameters passed to plot.window such as xlim, ylim, or mar.

Details

The device is a relatively recent feature of the package. It should support all operations but there might still be small inaccuracies. Also it is a bit slower than some of the other devices, in particular for rendering text and clipping. Hopefully this can be optimized in the next version.

By default image_draw sets all margins to 0 and uses graphics coordinates to match image size in pixels (width x height) where (0,0) is the top left corner. Note that this means the y axis increases from top to bottom which is the opposite of typical graphics coordinates. You can override all this by passing custom xlim, ylim or mar values to image_draw.

The image_capture function returns the current device as an image. This only works if the current device is a magick device or supports dev.capture.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# Regular image
frink <- image_read("https://jeroen.github.io/images/frink.png")

# Produce image using graphics device
fig <- image_graph(res = 96)
ggplot2::qplot(mpg, wt, data = mtcars, colour = cyl)
dev.off()

# Combine
out <- image_composite(fig, frink, offset = "+70+30")
print(out)

# Or paint over an existing image
img <- image_draw(frink)
rect(20, 20, 200, 100, border = "red", lty = "dashed", lwd = 5)
abline(h = 300, col = 'blue', lwd = '10', lty = "dotted")
text(10, 250, "Hoiven-Glaven", family = "monospace", cex = 4, srt = 90)
palette(rainbow(11, end = 0.9))
symbols(rep(200, 11), seq(0, 400, 40), circles = runif(11, 5, 35),
  bg = 1:11, inches = FALSE, add = TRUE)
dev.off()
print(img)


# Vectorized example with custom coordinates
earth <- image_read("https://jeroen.github.io/images/earth.gif")
img <- image_draw(earth, xlim = c(0,1), ylim = c(0,1))
rect(.1, .1, .9, .9, border = "red", lty = "dashed", lwd = 5)
text(.5, .9, "Our planet", cex = 3, col = "white")
dev.off()
print(img)

Edge / Line Detection

Description

Best results are obtained by finding edges with image_canny() and then performing Hough-line detection on the edge image.

Usage

image_edge(image, radius = 1)

image_canny(image, geometry = "0x1+10%+30%")

image_hough_draw(
  image,
  geometry = NULL,
  color = "red",
  bg = "transparent",
  size = 3,
  overlay = FALSE
)

image_hough_txt(image, geometry = NULL, format = c("mvg", "svg"))

Arguments

image

magick image object returned by image_read() or image_graph()

radius

edge size in pixels

geometry

geometry string, see details.

color

a valid color string such as "navyblue" or "#000080". Use "none" for transparency.

bg

background color

size

size in points to draw the line

overlay

composite the drawing atop the input image. Only for bg = 'transparent'.

format

output format of the text, either svg or mvg

Details

For Hough-line detection, the geometry format is ⁠{W}x{H}+{threshold}⁠ defining the size and threshold of the filter used to find 'peaks' in the intermediate search image. For canny edge detection the format is ⁠{radius}x{sigma}+{lower%}+{upper%}⁠. More details and examples are available at the imagemagick website.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

if(magick_config()$version > "6.8.9"){
shape <- demo_image("shape_rectangle.gif")
rectangle <- image_canny(shape)
rectangle |> image_hough_draw('5x5+20')
rectangle |> image_hough_txt(format = 'svg') |> cat()
}

Image Editing

Description

Read, write and join or combine images. All image functions are vectorized, meaning they operate either on a single frame or a series of frames (e.g. a collage, video, or animation). Besides paths and URLs, image_read() supports commonly used bitmap and raster object types.

Usage

image_read(
  path,
  density = NULL,
  depth = NULL,
  strip = FALSE,
  coalesce = TRUE,
  defines = NULL
)

image_read_svg(path, width = NULL, height = NULL)

image_read_pdf(path, pages = NULL, density = 300, password = "")

image_read_video(path, fps = 1, format = "png")

image_write(
  image,
  path = NULL,
  format = NULL,
  quality = NULL,
  depth = NULL,
  density = NULL,
  comment = NULL,
  flatten = FALSE,
  defines = NULL,
  compression = NULL
)

image_convert(
  image,
  format = NULL,
  type = NULL,
  colorspace = NULL,
  depth = NULL,
  antialias = NULL,
  matte = NULL,
  interlace = NULL,
  profile = NULL
)

image_data(image, channels = NULL, frame = 1)

image_raster(image, frame = 1, tidy = TRUE)

image_display(image, animate = TRUE)

image_browse(image, browser = getOption("browser"))

image_strip(image)

image_blank(width, height, color = "none", pseudo_image = "", defines = NULL)

image_destroy(image)

image_join(...)

image_attributes(image)

image_get_artifact(image, artifact = "")

demo_image(path)

Arguments

path

a file, url, or raster object or bitmap array

density

resolution to render pdf or svg

depth

color depth (either 8 or 16)

strip

drop image comments and metadata

coalesce

automatically image_coalesce() gif images

defines

a named character vector with extra options to control reading. These are the ⁠-define key{=value}⁠ settings in the command line tool. Use an empty string for value-less defines, and NA to unset a define.

width

in pixels

height

in pixels

pages

integer vector with page numbers. Defaults to all pages.

password

user password to open protected pdf files

fps

how many images to capture per second of video. Set to NULL to get all frames from the input video.

format

output format such as "png", "jpeg", "gif", "rgb" or "rgba".

image

magick image object returned by image_read() or image_graph()

quality

number between 0 and 100 for jpeg quality. Defaults to 75.

comment

text string added to the image metadata for supported formats

flatten

should image be flattened before writing? This also replaces transparency with background color.

compression

a string with compression type from compress_types

type

string with imagetype value from image_types for example grayscale to convert into black/white

colorspace

string with a colorspace from colorspace_types for example "gray", "rgb" or "cmyk"

antialias

enable anti-aliasing for text and strokes

matte

set to TRUE or FALSE to enable or disable transparency

interlace

string with interlace

profile

path to file with ICC color profile

channels

string with image channel(s) for example "rgb", "rgba", "cmyk","gray", or "ycbcr". Default is either "gray", "rgb" or "rgba" depending on the image

frame

integer setting which frame to extract from the image

tidy

converts raster data to long form for use with geom_raster. If FALSE output is the same as as.raster().

animate

support animations in the X11 display

browser

argument passed to browseURL

color

a valid color string such as "navyblue" or "#000080". Use "none" for transparency.

pseudo_image

string with pseudo image specification for example "radial-gradient:purple-yellow"

...

several images or lists of images to be combined

artifact

string with name of the artifact to extract, see the image_deskew for an example.

Details

All standard base vector methods such as [, [[, c(), as.list(), as.raster(), rev(), length(), and print() can be used to work with magick image objects. Use the standard img[i] syntax to extract a subset of the frames from an image. The img[[i]] method is an alias for image_data() which extracts a single frame as a raw bitmap matrix with pixel values.

For reading svg or pdf it is recommended to use image_read_svg() and image_read_pdf() if the rsvg and pdftools R packages are available. These functions provide more rendering options (including rendering of literal svg) and better quality than built-in svg/pdf rendering delegates from imagemagick itself.

X11 is required for image_display() which is only works on some platforms. A more portable method is image_browse() which opens the image in a browser. RStudio has an embedded viewer that does this automatically which is quite nice.

Image objects are automatically released by the garbage collector when they are no longer reachable. Because the GC only runs once in a while, you can also call image_destroy() explicitly to release the memory immediately. This is usually only needed if you create a lot of images in a short period of time, and you might run out of memory.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# Download image from the web
frink <- image_read("https://jeroen.github.io/images/frink.png")
worldcup_frink <- image_fill(frink, "orange", "+100+200", 20)
image_write(worldcup_frink, "output.png")

# extract raw bitmap array
bitmap <- frink[[1]]

# replace pixels with #FF69B4 ('hot pink') and convert back to image
bitmap[,50:100, 50:100] <- as.raw(c(0xff, 0x69, 0xb4, 0xff))
image_read(bitmap)

# Plot to graphics device via legacy raster format
raster <- as.raster(frink)
par(ask=FALSE)
plot(raster)

# Read bitmap arrays from other image packages
download.file("https://jeroen.github.io/images/example.webp", "example.webp", mode = 'wb')
if(require(webp)) image_read(webp::read_webp("example.webp"))
unlink(c("example.webp", "output.png"))
if(require(rsvg)){
tiger <- image_read_svg("http://jeroen.github.io/images/tiger.svg")
svgtxt <- '<?xml version="1.0" encoding="UTF-8"?>
<svg width="400" height="400" viewBox="0 0 400 400" fill="none">
 <circle fill="steelblue" cx="200" cy="200" r="100" />
 <circle fill="yellow" cx="200" cy="200" r="90" />
</svg>'
circles <- image_read_svg(svgtxt)
}
if(require(pdftools))
image_read_pdf(file.path(R.home('doc'), 'NEWS.pdf'), pages = 1, density = 100)
# create a solid canvas
image_blank(600, 400, "green")
image_blank(600, 400, pseudo_image = "radial-gradient:purple-yellow")
image_blank(200, 200, pseudo_image = "gradient:#3498db-#db3a34",
  defines = c('gradient:direction' = 'east'))

Image Effects

Description

High level effects applied to an entire image. These are mostly just for fun.

Usage

image_despeckle(image, times = 1L)

image_reducenoise(image, radius = 1L)

image_noise(image, noisetype = "gaussian")

image_blur(image, radius = 1, sigma = 0.5)

image_motion_blur(image, radius = 1, sigma = 0.5, angle = 0)

image_charcoal(image, radius = 1, sigma = 0.5)

image_oilpaint(image, radius = 1)

image_emboss(image, radius = 1, sigma = 0.5)

image_implode(image, factor = 0.5)

image_negate(image)

Arguments

image

magick image object returned by image_read() or image_graph()

times

number of times to repeat the despeckle operation

radius

radius, in pixels, for various transformations

noisetype

string with a noisetype value from noise_types.

sigma

the standard deviation of the Laplacian, in pixels.

angle

angle, in degrees, for various transformations

factor

image implode factor (special effect)

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, fx, geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

logo <- image_read("logo:")
image_despeckle(logo)
image_reducenoise(logo)
image_noise(logo)
image_blur(logo, 10, 10)
image_motion_blur(logo, 10, 10, 45)
image_charcoal(logo)
image_oilpaint(logo, radius = 3)
image_emboss(logo)
image_implode(logo)
image_negate(logo)

Image FX

Description

Apply a custom an fx expression to the image.

Usage

image_fx(image, expression = "p", channel = NULL)

image_fx_sequence(image, expression = "p")

Arguments

image

magick image object returned by image_read() or image_graph()

expression

string with an fx expression

channel

a value of channel_types() specifying which channel(s) to set

Details

There are two different interfaces. The image_fx function simply applies the same fx to each frame in the input image. The image_fx_sequence function on the other hand treats the entire input vector as a sequence, allowing you to apply an expression with multiple input images. See examples.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), geometry, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# Show image_fx() expression
img <- image_convert(logo, colorspace = "Gray")
gradient_x <- image_convolve(img, kernel = "Prewitt")
gradient_y <- image_convolve(img, kernel = "Prewitt:90")
gradient <- c(image_fx(gradient_x, expression = "p^2"),
                image_fx(gradient_y, expression = "p^2"))
gradient <- image_flatten(gradient, operator = "Plus")
#gradient <- image_fx(gradient, expression = "sqrt(p)")
gradient


image_fx(img, expression = "pow(p, 0.5)")
image_fx(img, expression = "rand()")

# Use multiple source images

input <- c(logo, image_flop(logo))
image_fx_sequence(input, "(u+v)/2")

Geometry Helpers

Description

ImageMagick uses a handy geometry syntax to specify coordinates and shapes for use in image transformations. You can either specify these manually as strings or use the helper functions below.

Usage

geometry_point(x, y)

geometry_area(width = NULL, height = NULL, x_off = 0, y_off = 0)

geometry_size_pixels(width = NULL, height = NULL, preserve_aspect = TRUE)

geometry_size_percent(width = 100, height = NULL)

Arguments

x

left offset in pixels

y

top offset in pixels

width

in pixels

height

in pixels

x_off

offset in pixels on x axis

y_off

offset in pixels on y axis

preserve_aspect

if FALSE, resize to width and height exactly, loosing original aspect ratio. Only one of percent and preserve_aspect may be TRUE.

Details

See ImageMagick Manual for details about the syntax specification. Examples of geometry strings:

  • "500x300"Resize image keeping aspect ratio, such that width does not exceed 500 and the height does not exceed 300.

  • "500x300!"Resize image to 500 by 300, ignoring aspect ratio

  • "500x"Resize width to 500 keep aspect ratio

  • "x300"Resize height to 300 keep aspect ratio

  • "50%x20%"Resize width to 50 percent and height to 20 percent of original

  • "500x300+10+20"Crop image to 500 by 300 at position 10,20

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, morphology, ocr, options(), painting, segmentation, transform(), video

Examples

# Specify a point
logo <- image_read("logo:")
image_annotate(logo, "Some text", location = geometry_point(100, 200), size = 24)

# Specify image area
image_crop(logo, geometry_area(300, 300), repage = FALSE)
image_crop(logo, geometry_area(300, 300, 100, 100), repage = FALSE)

# Specify image size
image_resize(logo, geometry_size_pixels(300))
image_resize(logo, geometry_size_pixels(height = 300))
image_resize(logo, geometry_size_pixels(300, 300, preserve_aspect = FALSE))

# resize relative to current size
image_resize(logo, geometry_size_percent(50))
image_resize(logo, geometry_size_percent(50, 20))

Image to ggplot

Description

Create a ggplot with axes set to pixel coordinates and plot the raster image on it using ggplot2::annotation_raster. See examples for how to plot an image onto an existing ggplot.

Usage

image_ggplot(image, interpolate = FALSE)

Arguments

image

magick image object returned by image_read() or image_graph()

interpolate

passed to ggplot2::annotation_raster

Examples

# Plot with base R
plot(logo)

# Plot image with ggplot2
library(ggplot2)
myplot <- image_ggplot(logo)
myplot + ggtitle("Test plot")

# Show that coordinates are reversed:
myplot + theme_classic()

# Or add to plot as annotation
image <- image_fill(logo, 'none')
raster <- as.raster(image)
myplot <- qplot(mpg, wt, data = mtcars)
myplot + annotation_raster(raster, 25, 35, 3, 5)

# Or overplot image using grid
library(grid)
qplot(speed, dist, data = cars, geom = c("point", "smooth"))
grid.raster(image)

Morphology

Description

Apply a morphology method. This is a very flexible function which can be used to apply any morphology method with custom parameters. See imagemagick website for examples.

Usage

image_morphology(
  image,
  method = "convolve",
  kernel = "Gaussian",
  iterations = 1,
  opts = list()
)

image_convolve(
  image,
  kernel = "Gaussian",
  iterations = 1,
  scaling = NULL,
  bias = NULL
)

Arguments

image

magick image object returned by image_read() or image_graph()

method

a string with a valid method from morphology_types()

kernel

either a square matrix or a string. The string can either be a parameterized kerneltype such as: "DoG:0,0,2" or "Diamond" or it can contain a custom matrix (see examples)

iterations

number of iterations

opts

a named list or character vector with custom attributes

scaling

string with kernel scaling. The special flag "!" automatically scales to full dynamic range, for example: "50%!"

bias

output bias string, for example "50%"

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, ocr, options(), painting, segmentation, transform(), video

Examples

#example from IM website:
if(magick_config()$version > "6.8.8"){
pixel <- image_blank(1, 1, 'white') |> image_border('black', '5x5')

# See the effect of Dilate method
pixel |> image_scale('800%')
pixel |> image_morphology('Dilate', "Diamond") |> image_scale('800%')

# These produce the same output:
pixel |> image_morphology('Dilate', "Diamond", iter = 3) |> image_scale('800%')
pixel |> image_morphology('Dilate', "Diamond:3") |> image_scale('800%')

# Plus example
pixel |> image_morphology('Dilate', "Plus", iterations = 2) |> image_scale('800%')

# Rose examples
rose |> image_morphology('ErodeI', 'Octagon', iter = 3)
rose |> image_morphology('DilateI', 'Octagon', iter = 3)
rose |> image_morphology('OpenI', 'Octagon', iter = 3)
rose |> image_morphology('CloseI', 'Octagon', iter = 3)

# Edge detection
man <- demo_image('man.gif')
man |> image_morphology('EdgeIn', 'Octagon')
man |> image_morphology('EdgeOut', 'Octagon')
man |> image_morphology('Edge', 'Octagon')

# Octagonal Convex Hull
 man |>
   image_morphology('Close', 'Diamond') |>
   image_morphology('Thicken', 'ConvexHull', iterations = 1)

# Thinning down to a Skeleton
man |> image_morphology('Thinning', 'Skeleton', iterations = 1)

# Specify custom kernel matrix usingn a string:
img <- demo_image("test_mag.gif")
i <- image_convolve(img, kernel = '4x5:
       0 -1  0  0
      -1 +1 -1  0
      -1 +1 -1  0
      -1 +1 +1 -1
       0 -1 -1  0 ', bias = "50%")
}

Image Text OCR

Description

Extract text from an image using the tesseract package.

Usage

image_ocr(image, language = "eng", HOCR = FALSE, ...)

image_ocr_data(image, language = "eng", ...)

Arguments

image

magick image object returned by image_read() or image_graph()

language

passed to tesseract. To install additional languages see instructions in tesseract_download().

HOCR

if TRUE return results as HOCR xml instead of plain text

...

additional parameters passed to tesseract

Details

To use this function you need to tesseract first:

  install.packages("tesseract")

Best results are obtained if you set the correct language in tesseract. To install additional languages see instructions in tesseract_download().

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, options(), painting, segmentation, transform(), video

Examples

if(require("tesseract")){
img <- image_read("http://jeroen.github.io/images/testocr.png")
image_ocr(img)
image_ocr_data(img)
}

Magick Options

Description

List option types and values supported in your version of ImageMagick. For descriptions see ImageMagick Enumerations.

Usage

magick_options()

magick_fonts()

option_types()

filter_types()

metric_types()

dispose_types()

compose_types()

colorspace_types()

channel_types()

image_types()

kernel_types()

noise_types()

gravity_types()

orientation_types()

morphology_types()

style_types()

decoration_types()

compress_types()

distort_types()

dump_option_info(option = "font")

Arguments

option

one of the option_types

Details

The dump_option_info function is equivalent to calling convert -list [option] on the command line. It does not return anything, it only makes ImageMagick print stuff to the console, use only for debugging.

References

ImageMagick Manual: Enumerations

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, painting, segmentation, transform(), video


Image Painting

Description

The image_fill() function performs flood-fill by painting starting point and all neighboring pixels of approximately the same color. Annotate prints some text on the image.

Usage

image_fill(image, color, point = "+1+1", fuzz = 0, refcolor = NULL)

image_annotate(
  image,
  text,
  gravity = "northwest",
  location = "+0+0",
  degrees = 0,
  size = 10,
  font = "",
  style = "normal",
  weight = 400,
  kerning = 0,
  decoration = NULL,
  color = NULL,
  strokecolor = NULL,
  strokewidth = NULL,
  boxcolor = NULL
)

Arguments

image

magick image object returned by image_read() or image_graph()

color

a valid color string such as "navyblue" or "#000080". Use "none" for transparency.

point

a geometry_point string indicating the starting point of the flood-fill

fuzz

relative color distance (value between 0 and 100) to be considered similar in the filling algorithm

refcolor

if set, fuzz color distance will be measured against this color, not the color of the starting point. Any color (within fuzz color distance of the given refcolor), connected to starting point will be replaced with the color. If the pixel at the starting point does not itself match the given refcolor (according to fuzz) then no action will be taken.

text

character vector of length equal to 'image' or length 1

gravity

string with gravity value from gravity_types.

location

geometry string with location relative to gravity

degrees

rotates text around center point

size

font-size in pixels

font

string with font family such as "sans", "mono", "serif", "Times", "Helvetica", "Trebuchet", "Georgia", "Palatino" or "Comic Sans". See magick_fonts() for what is available.

style

value of style_types for example "italic"

weight

thickness of the font, 400 is normal and 700 is bold, see magick_fonts().

kerning

increases or decreases whitespace between letters

decoration

value of decoration_types for example "underline"

strokecolor

a color string adds a stroke (border around the text)

strokewidth

set the strokewidth of the border around the text

boxcolor

a color string for background color that annotation text is rendered on.

Details

Note that more sophisticated drawing mechanisms are available via the graphics device using image_draw.

Setting a font, weight, style only works if your imagemagick is compiled with fontconfig support.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), segmentation, transform(), video

Examples

logo <- image_read("logo:")
logo <- image_background(logo, 'white')
image_fill(logo, "pink", point = "+450+400")
image_fill(logo, "pink", point = "+450+400", fuzz = 25)
# Add some text to an image
image_annotate(logo, "This is a test")
image_annotate(logo, "CONFIDENTIAL", size = 50, color = "red", boxcolor = "pink",
 degrees = 30, location = "+100+100")

# Setting fonts requires fontconfig support (and that you have the font)
image_annotate(logo, "The quick brown fox", font = "monospace", size = 50)

Image Segmentation

Description

Basic image segmentation like connected components labelling, blob extraction and fuzzy c-means

Usage

image_connect(image, connectivity = 4)

image_split(image, keep_color = TRUE)

image_fuzzycmeans(image, min_pixels = 1, smoothing = 1.5)

Arguments

image

magick image object returned by image_read() or image_graph()

connectivity

number neighbor colors which are considered part of a unique object

keep_color

if TRUE the output images retain the color of the input pixel. If FALSE all matching pixels are set black to retain only the image mask.

min_pixels

the minimum number of pixels contained in a hexahedra before it can be considered valid (expressed as a percentage)

smoothing

the smoothing threshold which eliminates noise in the second derivative of the histogram (higher values gives smoother second derivative)

Details

  • image_connect Connect adjacent pixels with the same pixel intensities to do blob extraction

  • image_split Splits the image according to pixel intensities

  • image_fuzzycmeans Fuzzy c-means segmentation of the histogram of color components

image_connect performs blob extraction by scanning the image, pixel-by-pixel from top-left to bottom-right where regions of adjacent pixels which share the same set of intensity values get combined.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, transform(), video

Examples

# Split an image by color
img <- image_quantize(logo, 4)
layers <- image_split(img)
layers

# This returns the original image
image_flatten(layers)

# From the IM website
objects <- image_convert(demo_image("objects.gif"), colorspace = "Gray")
objects


# Split image in blobs of connected pixel levels
if(magick_config()$version > "6.9.0"){
objects |>
  image_connect(connectivity = 4) |>
  image_split()

# Fuzzy c-means
image_fuzzycmeans(logo)

logo |>
  image_convert(colorspace = "HCL") |>
  image_fuzzycmeans(smoothing = 5)
}

Image thresholding

Description

Thresholding an image can be used for simple and straightforward image segmentation. The function image_threshold() allows to do black and white thresholding whereas image_lat() performs local adaptive thresholding.

Usage

image_threshold(
  image,
  type = c("black", "white"),
  threshold = "50%",
  channel = NULL
)

image_level(
  image,
  black_point = 0,
  white_point = 100,
  mid_point = 1,
  channel = NULL
)

image_lat(image, geometry = "10x10+5%")

Arguments

image

magick image object returned by image_read() or image_graph()

type

type of thresholding, either one of lat, black or white (see details below)

threshold

pixel intensity threshold percentage for black or white thresholding

channel

a value of channel_types() specifying which channel(s) to set

black_point

value between 0 and 100, the darkest color in the image

white_point

value between 0 and 100, the lightest color in the image

mid_point

value between 0 and 10 used for gamma correction

geometry

pixel window plus offset for LAT algorithm

Details

  • image_threshold(type = "black"): Forces all pixels below the threshold into black while leaving all pixels at or above the threshold unchanged

  • image_threshold(type = "white"): Forces all pixels above the threshold into white while leaving all pixels at or below the threshold unchanged

  • image_lat(): Local Adaptive Thresholding. Looks in a box (width x height) around the pixel neighborhood if the pixel value is bigger than the average minus an offset.

Examples

test <- image_convert(logo, colorspace = "Gray")
image_threshold(test, type = "black", threshold = "50%")
image_threshold(test, type = "white", threshold = "50%")

# Turn image into BW
test |>
  image_threshold(type = "white", threshold = "50%") |>
  image_threshold(type = "black", threshold = "50%")

# adaptive thresholding
image_lat(test, geometry = '10x10+5%')

Image Transform

Description

Basic transformations like rotate, resize, crop and flip. The geometry syntax is used to specify sizes and areas.

Usage

image_trim(image, fuzz = 0)

image_chop(image, geometry)

image_rotate(image, degrees)

image_resize(image, geometry = NULL, filter = NULL)

image_scale(image, geometry = NULL)

image_sample(image, geometry = NULL)

image_crop(image, geometry = NULL, gravity = NULL, repage = TRUE)

image_extent(image, geometry, gravity = "center", color = "none")

image_flip(image)

image_flop(image)

image_deskew(image, threshold = 40)

image_deskew_angle(image, threshold = 40)

image_page(image, pagesize = NULL, density = NULL)

image_repage(image)

image_orient(image, orientation = NULL)

image_shear(image, geometry = "10x10", color = "none")

image_distort(image, distortion = "perspective", coordinates, bestfit = FALSE)

Arguments

image

magick image object returned by image_read() or image_graph()

fuzz

relative color distance (value between 0 and 100) to be considered similar in the filling algorithm

geometry

a geometry string specifying area (for cropping) or size (for resizing).

degrees

value between 0 and 360 for how many degrees to rotate

filter

string with filter type from: filter_types

gravity

string with gravity value from gravity_types.

repage

resize the canvas to the cropped area

color

a valid color string such as "navyblue" or "#000080". Use "none" for transparency.

threshold

straightens an image. A threshold of 40 works for most images.

pagesize

geometry string with preferred size and location of an image canvas

density

geometry string with vertical and horizontal resolution in pixels of the image. Specifies an image density when decoding a Postscript or PDF.

orientation

string to set image orientation one of the orientation_types. If NULL it applies auto-orientation which tries to infer the correct orientation from the Exif data.

distortion

string to set image orientation one of the distort_types.

coordinates

numeric vector (typically of length 12) with distortion coordinates

bestfit

if set to TRUE the size of the output image can be different from input

Details

For details see Magick++ STL documentation. Short descriptions:

The most powerful resize function is image_resize which allows for setting a custom resize filter. Output of image_scale is similar to image_resize(img, filter = "point").

For resize operations it holds that if no geometry is specified, all frames are rescaled to match the top frame.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, video

Examples

logo <- image_read("logo:")
logo <- image_scale(logo, "400")
image_trim(logo)
image_chop(logo, "100x20")
image_rotate(logo, 45)
# Small image
rose <- image_convert(image_read("rose:"), "png")

# Resize to 400 width or height:
image_resize(rose, "400x")
image_resize(rose, "x400")

# Resize keeping ratio
image_resize(rose, "400x400")

# Resize, force size losing ratio
image_resize(rose, "400x400!")

# Different filters
image_resize(rose, "400x", filter = "Triangle")
image_resize(rose, "400x", filter = "Point")
# simple pixel resize
image_scale(rose, "400x")
image_sample(rose, "400x")
image_crop(logo, "400x400+200+200")
image_extent(rose, '200x200', color = 'pink')
image_flip(logo)
image_flop(logo)
skewed <- image_rotate(logo, 5)
deskewed <- image_deskew(skewed)
attr(deskewed, 'angle')
if(magick_config()$version > "6.8.6")
  image_orient(logo)
image_shear(logo, "10x10")
building <- demo_image('building.jpg')
image_distort(building, 'perspective', c(7,40,4,30,4,124,4,123,85,122,100,123,85,2,100,30))

Write Video

Description

High quality video / gif exporter based on external packages gifski and av.

Usage

image_write_video(image, path = NULL, framerate = 10, ...)

image_write_gif(image, path = NULL, delay = 1/10, ...)

Arguments

image

magick image object returned by image_read() or image_graph()

path

filename of the output gif or video. This is also the return value.

framerate

frames per second, passed to av_encode_video

...

additional parameters passed to av_encode_video and gifski.

delay

duration of each frame in seconds (inverse of framerate)

Details

This requires an image with multiple frames. The GIF exporter accomplishes the same thing as image_animate but much faster and with better quality.

See Also

Other image: _index_, analysis, animation, attributes(), color, composite, defines, device, edges, editing, effects(), fx, geometry, morphology, ocr, options(), painting, segmentation, transform()


Example Images

Description

Example images included with ImageMagick:

Usage

logo

Format

An object of class magick-image of length 1.

Details

  • logo: ImageMagick Logo, 640x480

  • wizard: ImageMagick Wizard, 480x640

  • rose : Picture of a rose, 70x46

  • granite : Granite texture pattern, 128x128