Documentation
¶
Overview ¶
Package sgrad provides facilities for the computation of a finite-difference gradient image from a source SippImage and a 2x2 kernel. There are two versions, one using float64s and complex128s, and another using int32s. The latter makes it easier to guarantee bit accuracy and numerical stability. It is not intended as a performance optimisation.
Index ¶
- func Fdgrad(src SippImage) (grad *ComplexImage)
- func FdgradInt32(src SippImage) (grad *ComplexInt32Image)
- func FdgradInt32Kernel(src SippImage, kern SippGradInt32Kernel) (grad *ComplexInt32Image)
- func FdgradKernel(src SippImage, kern SippGradKernel) (grad *ComplexImage)
- type SippGradInt32Kernel
- type SippGradKernel
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func Fdgrad ¶
func Fdgrad(src SippImage) (grad *ComplexImage)
Use a default SippGradKernel to compute a finite-differences gradient. See FdgradKernel for details.
func FdgradInt32 ¶
func FdgradInt32(src SippImage) (grad *ComplexInt32Image)
Use a default SippGradInt32Kernel to compute a finite-differences gradient. See FdgradInt32Kernel for details.
func FdgradInt32Kernel ¶
func FdgradInt32Kernel(src SippImage, kern SippGradInt32Kernel) (grad *ComplexInt32Image)
Use a SippGradInt32Kernel to create a finite-differences ComplexInt32 gradient image, one pixel narrower and shorter than the original. We'd rather reduce the size of the output image than arbitrarily wrap around or extend the source image, as any such procedure could introduce errors into the statistics.
func FdgradKernel ¶
func FdgradKernel(src SippImage, kern SippGradKernel) (grad *ComplexImage)
Use a SippGradKernel to create a finite-differences complex gradient image, one pixel narrower and shorter than the original. We'd rather reduce the size of the output image than arbitrarily wrap around or extend the source image, as any such procedure could introduce errors into the statistics.
Types ¶
type SippGradInt32Kernel ¶
type SippGradInt32Kernel [2][2]ComplexInt32
type SippGradKernel ¶
type SippGradKernel [2][2]complex128
SippGradKernels are 2x2 arrays of complex numbers, defined in the same way as images are stored in memory, i.e. in row-major order from the top-left corner down.
Source Files