agabor

Documentation

Index

• func Convolve(ch int, segmentSteps int, borderSteps int, melFilterCount int, ...)
• func ToTensor(specs []Filter, set *FilterSet)
• type Filter
  • func (f *Filter) Defaults()
• type FilterSet

Functions

func Convolve

func Convolve(ch int, segmentSteps int, borderSteps int, melFilterCount int, melData *etensor.Float32, filters FilterSet, rawOut *etensor.Float32)

Convolve processes input using filters that operate over an entire segment of samples

func ToTensor

func ToTensor(specs []Filter, set *FilterSet)

ToTensor generates filters into the tensor passed by caller

Types

type Filter

type Filter struct {
	SizeX       int     `` /* 138-byte string literal not displayed */
	SizeY       int     `` /* 159-byte string literal not displayed */
	WaveLen     float32 `desc:"wavelength of the sine waves in normalized units, 1.5 and 2 are reasonable values"`
	Orientation float32 `` /* 148-byte string literal not displayed */
	SigmaLength float32 `` /* 170-byte string literal not displayed */
	SigmaWidth  float32 `` /* 157-byte string literal not displayed */
	PhaseOffset float32 `` /* 146-byte string literal not displayed */
	CircleEdge  bool    `` /* 154-byte string literal not displayed */
}

Filter, a struct of gabor filter parameters

func (*Filter) Defaults

func (f *Filter) Defaults()

Defaults sets default values for any filter fields where 0 is not a reasonable value

type FilterSet

type FilterSet struct {
	SizeX   int             `desc:"size of each filter in X"`
	SizeY   int             `desc:"size of each filter in Y"`
	StrideX int             `desc:"how far to move the filter in X each step"`
	StrideY int             `desc:"how far to move the filter in Y each step"`
	Gain    float32         `` /* 145-byte string literal not displayed */
	Filters etensor.Float32 `desc:"actual gabor filters"`
}

FilterSet, a struct holding a set of gabor filters stored as a tensor. Though individual filters can vary in size, when used as a set they should all have the same size.