pub trait KernelGenerator {
// Required methods
fn required_number_of_points(&self) -> usize;
fn get_kernel_value(&self, offset: f64) -> f64;
// Provided methods
fn absolute_value_offset(&self) -> usize { ... }
fn expected_weighted_sum(&self) -> f64 { ... }
}
Available on crate feature
std
only.Expand description
A convolution kernel generator.
Required Methods§
Sourcefn required_number_of_points(&self) -> usize
fn required_number_of_points(&self) -> usize
The total number of points required for this kernel. Cannot be zero - usually an odd value.
Sourcefn get_kernel_value(&self, offset: f64) -> f64
fn get_kernel_value(&self, offset: f64) -> f64
Calculates the value for the convolution kernel at the specified offset. The range of valid offsets is [-offset_size, 0, offset_size]
Provided Methods§
Sourcefn absolute_value_offset(&self) -> usize
fn absolute_value_offset(&self) -> usize
The minimum (negative) and maximum (positive) index into the [get_kernel_value
] function. This defaults to [(1-N)/2, (N-1)/2].
Sourcefn expected_weighted_sum(&self) -> f64
fn expected_weighted_sum(&self) -> f64
Returns the expected value of the kernel. Most kernels will be 1.0 though some (like first-derivative kernels) will be at zero.