equalizeHistogram image enhancement via colour histogram equalization.
Usage
equalize_image(
vesalius_assay,
dimensions = seq(1, 3),
embedding = "last",
method = "BalanceSimplest",
N = 1,
smax = 1,
sleft = 1,
sright = 1,
lambda = 0.1,
up = 100,
down = 10,
verbose = TRUE
)Arguments
- vesalius_assay
a vesalius_assay object
- dimensions
numeric vector of latent space dimensions to use.
- embedding
character string describing which embedding should be used.
- N
numeric describing how each colour channel will be mapped back to the image (Higher N = Higher greyscale contrast). Used with EqualizePiecewise
- smax
numeric - upper limit if contrast stretching. Used with EqualizePiecewise
- sleft
numeric - Range 0 - 100. Percentage of pixel to be saturated on the left side of the histogram. Used with BalanceSimplest
- sright
numeric - Range 0 - 100. Percentage of pixel to be saturated on the right side of the histogram. Used with BalanceSimplest
- lambda
numeric - strength of background correction. Used with SPE (Screened Poisson Equation).
- up
numeric - color value threshold in the upper limit. Used with EqualizeDP.
- down
numeric color value threshold in the lower limit. Used with EqualizeDP.
- verbose
logical - progress message output.
- type
character - histogram EQ type. Select from: BalanceSimplest, EqualizePiecewise, SPE, EqualizeDP, EqualizeADP, ECDF (see details)
Details
Histogram equalization ensures that image details are amplified. In turn, territories may be extract with greater precision. We recommend balancing the histogram prior to smoothing.
For further details on each method described here, please refer to https://cran.r-project.org/web/packages/imagerExtra/vignettes/gettingstarted.html
Examples
if (FALSE) { # \dontrun{
data(vesalius)
# First we build a simple object
ves <- build_vesalius_object(coordinates, counts)
# We can do a simple run
ves <- build_vesalius_embeddings(ves)
# simple EQ
ves <- equalisz_image(ves, embedding = "PCA")
} # }