Closed dlebauer closed 7 years ago
In addition, Following the function examples, setting this para.meter to different values does not change the output. For example:
library(EBImage)
example('computeFeatures')
all(computeFeatures.haralick(x, y, scales = 1) == computeFeatures.haralick(x, y, scales = c(1,2)))
##TRUE
identical(computeFeatures.haralick(x, y, scales = 1), computeFeatures.haralick(x, y, scales = c(2,6)))
##TRUE
The question has been already answered by @kevin05jan.
The reason why you were getting identical results for seemingly different function calls was due to wrong argument name: scales
instead of haralick.scales
.
In the documentation, the function
computeFeatures
has a parameterharalick.scales
with a default value ofc(1, 2)
. The parameter is defined as:However, it is not clear what this refers to, even after a brief look at the code https://github.com/aoles/EBImage/blob/master/R/computeFeatures.R#L248.
In the original Haralick paper (Haralick 1973.pdf), the author refers to using both the mean and range of each feature. However, it is not clear if this is related or even an optional output from the algorithm.