This function performs a binary "not" operation on all bits of the binary representation of an input pixel's gray value. A 0-bit in the input pixel results in an 1-bit in the corresponding output pixel, an 1-bit in the input gets a 0-bit in the output. For integral image types, the resulting output gray value is the difference of the maximum possible gray value minus the input gray value, but for float image type the results are strange due to the inhomogeneous float format.