Module |
|
Header |
/Engine/Source/Runtime/SignalProcessing/Public/DSP/FloatArrayMath.h |
Include |
#include "DSP/FloatArrayMath.h" |
class FContiguousSparse2DKernelTransform
FContiguousSparse2DKernelTransform
FContiguousSparse2DKernelTransform applies a matrix transformation to an input array. [OutArray] = [[Kernal]][InView]
It provides some optimization by exploit the contiguous and sparse qualities of the kernel rows, which allows it to skip multiplications with the number zero.
It works with non-sparse and non-contiguous kernels as well, but will be more computationally expensive than a naive implementation. Also, only takes advantage of sparse contiguous rows, not columns.
Name | Description | |
---|---|---|
|
FContiguousSparse2DKernelTransform ( |
|
|
FContiguousSparse2DKernelTransform ( |
|
|
FContiguousSparse2DKernelTransform |
NumInElements sets the expected number of input array elements as well as the number of elements in a row. |
Name | Description | |
---|---|---|
|
~FContiguousSparse2DKernelTransform() |
Name | Description | ||
---|---|---|---|
|
GetNumInElements() |
Returns the required size of the input array |
|
|
GetNumOutElements() |
Returns the size of the output array |
|
|
SetRow ( |
Set the kernel values for an individual row. |
|
|
TransformArray ( |
Transforms the input array given the kernel. |
|
|
TransformArray ( |
Transforms the input array given the kernel. |
|
|
TransformArray ( |
Transforms the input array given the kernel. |
Name | Description | ||
---|---|---|---|
|
FContiguousS... |
operator= ( |