| Title: | Apply Photo Editing Effects |
|---|---|
| Description: | You can apply image processing effects that modifies the perceived material properties of objects in photos, such as gloss, smoothness, and blemishes. This is an implementation of the algorithm proposed by Boyadzhiev et al. (2015) "Band-Sifting Decomposition for Image Based Material Editing". Documentation and practical tips of the package is available at <https://github.com/tsuda16k/materialmodifier>. |
| Authors: | Hiroyuki Tsuda [aut, cre] |
| Maintainer: | Hiroyuki Tsuda <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 1.2.0 |
| Built: | 2025-02-25 04:15:36 UTC |
| Source: | https://github.com/tsuda16k/materialmodifier |
cimg to nimg conversion
cimg2nimg(im)cimg2nimg(im)
im |
a cimg object |
an nimg object
A photograph obtained from a free stock photos site. pexels.com/photo/fashion-woman-cute-shoes-5704849/
faceface
An array with 500 x 500 x 3 dimensions. Each dimension represents y-coordinate, x-coordinate, and color channel.
plot(face)plot(face)
Calculate the BS feature energy
get_BS_energy(im, mask = NA, logspace = TRUE)get_BS_energy(im, mask = NA, logspace = TRUE)
im |
An image. |
mask |
(optional) If set, only the area of white pixels in the mask image will be included in the calculation. |
logspace |
If TRUE (default), image processing is done in the log space. If FALSE, computation is performed without log transformation. |
a data frame
## Not run: data = get_BS_energy(face) ## End(Not run)## Not run: data = get_BS_energy(face) ## End(Not run)
Scale-space decomposition by the guided filter
gf_decompose( im, mask = NA, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )gf_decompose( im, mask = NA, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )
im |
an image |
mask |
If set, only the area of white pixels in the mask image will be edited. |
log_epsilon |
offset for log transformation |
filter_epsilon |
epsilon parameter |
logspace |
If TRUE (default), image processing is done in the log space. If FALSE, computation is performed without log transformation. |
a list of images
Scale-space decomposition
gf_decompose_parts(dec, mask = NA)gf_decompose_parts(dec, mask = NA)
dec |
output of gf_decompose_scale function |
mask |
If set, only the area of white pixels in the mask image will be edited. |
a list of images
Scale-space decomposition by the guided filter
gf_decompose_scale( im, depth = NULL, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )gf_decompose_scale( im, depth = NULL, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )
im |
a grayscale image |
depth |
scale depth |
log_epsilon |
offset for log transformation |
filter_epsilon |
epsilon parameter |
logspace |
If TRUE (default), image processing is done in the log space. If FALSE, computation is performed without log transformation. |
a list of images
Reconstruct the original image from decomposed data
gf_reconstruct(dec, scales, ind, include.residual = TRUE, logspace = TRUE)gf_reconstruct(dec, scales, ind, include.residual = TRUE, logspace = TRUE)
dec |
decomposed data |
scales |
which spatial scales to use for reconstruction |
ind |
a numeric vector |
include.residual |
either TRUE (default) or FALSE |
logspace |
If TRUE (default), image processing is done in the log space. If FALSE, computation is performed without log transformation. |
an image
Load image from file or URL
im_load(file, name)im_load(file, name)
file |
path to file or URL |
name |
a string for name attribute. if missing, inferred from the file argument. |
an array of image data
## Not run: # load an image from disk im = im_load("path/to/your/image.jpg") plot(im) ## End(Not run) # load an image from URL im = im_load("http://placehold.jp/150x150.png")## Not run: # load an image from disk im = im_load("path/to/your/image.jpg") plot(im) ## End(Not run) # load an image from URL im = im_load("http://placehold.jp/150x150.png")
Save an image to disk
im_save(im, name, path, format = "png", quality = 0.95)im_save(im, name, path, format = "png", quality = 0.95)
im |
An image. |
name |
Name of the image file. |
path |
The image is saved in this directory. For example, path = getwd() |
format |
Image format. Either "jpg", "png", "tiff", or "bmp". Default is "png". |
quality |
(jpg only) default is 0.95. Higher quality means less compression. |
No return value, called for side effects.
## Not run: # face.png is saved to a path (if a path is specified) im_save( face, path = "yourpath" ) # img.png is saved to a path (if a path is specified) im_save( face, name = "img", path = "yourpath" ) # myimage.jpg is saved to a path (if a path is specified) im_save( face, name = "myimage", path = "yourpath", format = "jpg" ) ## End(Not run)## Not run: # face.png is saved to a path (if a path is specified) im_save( face, path = "yourpath" ) # img.png is saved to a path (if a path is specified) im_save( face, name = "img", path = "yourpath" ) # myimage.jpg is saved to a path (if a path is specified) im_save( face, name = "myimage", path = "yourpath", format = "jpg" ) ## End(Not run)
This function is the core function of this package. It edits the input image by specifying the name of the editing effect (BS feature or its alias) and the strength parameter.
modif( im, effect, strength, mask = NA, max_size = 1280, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )modif( im, effect, strength, mask = NA, max_size = 1280, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )
im |
An input image. |
effect |
A string naming the effect to apply. Either "gloss", "shine", "spots", "blemish", "rough", "stain", "shadow", or "aging". |
strength |
A numeric, which controls the strength of the effect. Strength values between 0 and 1 will reduce a feature, while strength values larger than 1 will boost a feature. A strength value of 1 does nothing. Negative values are allowed, which will invert a feature. |
mask |
If set, only the area of white pixels in the mask image will be edited. |
max_size |
If the shorter side of the input image is larger than this value (the default is 1280), input image is resized before applying effects. Because the modif() function is very slow for large-resolution images, it is useful to limit the image resolution to speed-up the image processing. If you do not want to change the resolution of the input image, you can enter a large value for max_size, or set max_size = NA |
log_epsilon |
Offset for log transformation (default is 0.0001). Need not to change this value in most cases. |
filter_epsilon |
Epsilon parameter of the Guided filter (default is 0.01). Need not to change this value in most cases. |
logspace |
If TRUE (default), image processing is done in the log space. If FALSE, computation is performed without log transformation. |
an output image
plot(modif(face, effect = "shine", strength = 2.5)) # Apply the "shine" effect (make objects shiny) plot(modif(face, effect = "shine", strength = 0.2)) # Less shiny effect with a parameter less than 1 plot(modif(face, effect = c("shine", "stain"), strength = c(0.2, 3))) # Less shiny and more stainplot(modif(face, effect = "shine", strength = 2.5)) # Apply the "shine" effect (make objects shiny) plot(modif(face, effect = "shine", strength = 0.2)) # Less shiny effect with a parameter less than 1 plot(modif(face, effect = c("shine", "stain"), strength = c(0.2, 3))) # Less shiny and more stain
Check the scale information of an image
modif_dim(im)modif_dim(im)
im |
An image. |
A list of depth (number of scale subband images), indexes of high amplitude subbands, and indexes of low amplitude subbands.
modif_dim(face)modif_dim(face)
This function allows you to specify which image components to edit in more detail than the modif function. Please refer to the information on the package's Github page for detailed usage and theoretical background.
modif2( im, params, mask = NA, max_size = 1280, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )modif2( im, params, mask = NA, max_size = 1280, log_epsilon = 1e-04, filter_epsilon = 0.01, logspace = TRUE )
im |
An input image. |
params |
A list of parameter values. Parameter names are freq, amp, sign, and strength. |
mask |
If set, only the area of white pixels in the mask image will be edited. |
max_size |
If the shorter side of the input image is larger than this value (the default is 1280), input image is resized. The modif function is very slow for large-resolution images. |
log_epsilon |
Offset for log transformation (default is 0.0001). Need not to change this value in most cases. |
filter_epsilon |
Epsilon parameter of the Guided filter (default is 0.01). Need not to change this value in most cases. |
logspace |
If TRUE (default), image processing is done in the log space. If FALSE, computation is performed without log transformation. |
an output image
# shine effect shine = list(freq = "H", amp = "H", sign = "P", strength = 2) plot(modif2(face, params = shine)) # shine effect (equivalent to the above) shine2 = list(freq = 1:4, amp = "H", sign = "P", strength = 2) plot(modif2(face, params = shine2)) # you can specify a feature name directly, instead of specifying freq/amp/sign separately plot( modif2( face, params = list( feature = "HHA", strength = 2 ) ) ) plot( modif2( face, params = list( feature = "1HP", strength = 3 ) ) ) # apply multiple effects at the same time blemish = list(feature = "HLA", strength = 0.1) # less blemish smooth = list(feature = "HHN", strength = 0.2) # smoother plot(modif2(face, params = list(blemish, smooth)))# shine effect shine = list(freq = "H", amp = "H", sign = "P", strength = 2) plot(modif2(face, params = shine)) # shine effect (equivalent to the above) shine2 = list(freq = 1:4, amp = "H", sign = "P", strength = 2) plot(modif2(face, params = shine2)) # you can specify a feature name directly, instead of specifying freq/amp/sign separately plot( modif2( face, params = list( feature = "HHA", strength = 2 ) ) ) plot( modif2( face, params = list( feature = "1HP", strength = 3 ) ) ) # apply multiple effects at the same time blemish = list(feature = "HLA", strength = 0.1) # less blemish smooth = list(feature = "HHN", strength = 0.2) # smoother plot(modif2(face, params = list(blemish, smooth)))
nimg to cimg conversion
nimg2cimg(im)nimg2cimg(im)
im |
an nimg object |
a cimg object
Display an image
## S3 method for class 'nimg' plot(x, rescale = FALSE, ...)## S3 method for class 'nimg' plot(x, rescale = FALSE, ...)
x |
an image |
rescale |
logical. if true, then pixel value is rescaled to range between 0 and 1. |
... |
other parameters to be passed to plot.default |
No return value, called for side effects.
plot(face)plot(face)