Apply a kernel to the given histogram.
apply_kernel_histogram(
fhistogram,
filter = "circle",
mask = default_kernel(filter, radius, sigma),
radius = 2,
sigma = radius/2,
threads = 0
)
Matrix or array interpreted as histogram of floating-point values.
Use the pre-defined filter, either circle
, square
, gauss
. Defaults to circle
.
Custom kernel used for blurring, overrides filter
. Must be a square matrix of odd size.
Radius of the kernel (counted without the "middle" pixel"), defaults to 2. The generated kernel matrix will be a square with (2*radius+1) pixels on each side.
Radius of the Gaussian function selected by filter
, defaults to radius/2
.
Number of parallel threads (default 0 chooses hardware concurrency).
2D matrix with the histogram processed by the kernel application.