image

package

v0.0.0-...-1611311 Latest Latest Go to latest Published: Jul 15, 2023 License: Apache-2.0 Imports: 6 Imported by: 2

Details

Valid go.mod file

The Go module system was introduced in Go 1.11 and is the official dependency management solution for Go.
Redistributable license

Redistributable licenses place minimal restrictions on how software can be used, modified, and redistributed.
Tagged version

Modules with tagged versions give importers more predictable builds.
Stable version

When a project reaches major version v1 it is considered stable.
Learn more about best practices

Repository

github.com/galeone/tfgo

Links

Open Source Insights

Documentation ¶

Constants ¶

This section is empty.

Variables ¶

This section is empty.

Functions ¶

This section is empty.

Types ¶

type Box ¶

type Box struct {
	Start, End Point
}

Box represents the coordinates of 2 points in the space

type Image ¶

type Image struct {
	*tfgo.Tensor
}

Image is the type that abstracts a Tensorflow tensor representing an image. The image is always a float32 value unless manually casted.

func NewImage ¶

func NewImage(scope *op.Scope, tensor tf.Output) *Image

NewImage creates an *Image from a 3 or 4D input tensor Place the created image within the specified scope

func Read ¶

func Read(scope *op.Scope, imagePath string, channels int64) *Image

Read searches for the `imagePath` extensions and uses the Read<format> function to decode and load the right image. Panics if the format is unknown.

func ReadGIF ¶

func ReadGIF(scope *op.Scope, imagePath string) *Image

ReadGIF reads the GIF image whose path is `imagePath` and returns an Image

func ReadJPEG ¶

func ReadJPEG(scope *op.Scope, imagePath string, channels int64) *Image

ReadJPEG reads the JPEG image whose path is `imagePath` that has `channels` channels it returns an Image

func ReadPNG ¶

func ReadPNG(scope *op.Scope, imagePath string, channels int64) *Image

ReadPNG reads the PNG image whose path is `imagePath` that has `channels` channels it returns an Image

func (*Image) Add ¶

func (image *Image) Add(tfout tf.Output) *Image

Add defines the add operation between the image and tfout `tfout` dtype is converted to image.Dtype() before adding

func (*Image) AdjustBrightness ¶

func (image *Image) AdjustBrightness(delta float32) *Image

AdjustBrightness adds delta to the image

func (*Image) AdjustContrast ¶

func (image *Image) AdjustContrast(contrastFactor float32) *Image

AdjustContrast changes the contrast by contrastFactor

func (*Image) AdjustGamma ¶

func (image *Image) AdjustGamma(gamma, gain float32) *Image

AdjustGamma performs gamma correction on the image

func (*Image) AdjustHue ¶

func (image *Image) AdjustHue(delta float32) *Image

AdjustHue changes toe Hue by delta

func (*Image) AdjustSaturation ¶

func (image *Image) AdjustSaturation(saturationFactor float32) *Image

AdjustSaturation changes the saturation by saturationFactor

func (*Image) Cast ¶

func (image *Image) Cast(dtype tf.DataType) *Image

Cast casts the current image tensor to the requested type

func (*Image) Center ¶

func (image *Image) Center() *Image

Center computes the mean value of the pixel values and subtracts this value from every pixel: this operation centers the data

func (*Image) CentralCrop ¶

func (image *Image) CentralCrop(centralFraction float32) *Image

CentralCrop extracts from the center of the image a portion of image with an area equals to the centralFraction

func (*Image) Clone ¶

func (image *Image) Clone() *Image

Clone returns a copy of the current image in a new scope Clone must be used when one want to create a different image from the output of an operation.

func (*Image) ConvertDtype ¶

func (image *Image) ConvertDtype(dtype tf.DataType, saturate bool) *Image

ConvertDtype converts the Image dtype to dtype, uses SaturatesCast if required

func (*Image) Convolve ¶

func (image *Image) Convolve(filter tf.Output, stride Stride, padding padding.Padding) *Image

Convolve executes the convolution operation between the current image and the passed filter The strides parameter rules the stride along each dimension Padding is a padding type to specify the type of padding

func (*Image) Correlate ¶

func (image *Image) Correlate(filter tf.Output, stride Stride, padding padding.Padding) *Image

Correlate executes the correlation operation between the current image and the passed filter The strides parameter rules the stride along each dimension Padding is a padding type to specify the type of padding

func (*Image) CropAndResize ¶

func (image *Image) CropAndResize(box Box, size Size, optional ...op.CropAndResizeAttr) *Image

CropAndResize crops the image to the specified box and resizes the result to size

func (*Image) Dilate ¶

func (image *Image) Dilate(filter tf.Output, stride, rate Stride, padding padding.Padding) *Image

Dilate executes the dilatation operation between the current image and the padded filter The strides parameter rules the stride along each dimension, in output. The rate parameter rules the input stride for atrous morphological dilatation Padding is a padding type to specify the type of padding

func (*Image) DrawBoundingBoxes ¶

func (image *Image) DrawBoundingBoxes(boxes []Box) *Image

DrawBoundingBoxes draws the specified boxes to the image

func (*Image) EncodeJPEG ¶

func (image *Image) EncodeJPEG(optional ...op.EncodeJpegAttr) tf.Output

EncodeJPEG encodes the image in the JPEG format and returns an evaluable tensor

func (*Image) EncodePNG ¶

func (image *Image) EncodePNG(optional ...op.EncodePngAttr) tf.Output

EncodePNG encodes the image in the PNG format and returns an evaluable tensor

func (*Image) Erode ¶

func (image *Image) Erode(filter tf.Output, stride, rate Stride, padding padding.Padding) *Image

Erode executes the erosion operation between the current image and the padded filter The strides parameter rules the stride along each dimension The rate parameter rules the input stride for atrous morphological dilatation Padding is a padding type to specify the type of padding

func (*Image) ExtractGlimpse ¶

func (image *Image) ExtractGlimpse(size Size, offset Point, optional ...op.ExtractGlimpseAttr) tf.Output

ExtractGlimpse extracts a glimpse with the specified size at the specified offset

func (*Image) HSVToRGB ¶

func (image *Image) HSVToRGB() *Image

HSVToRGB performs the colorspace transformation from HSV to RGB

func (*Image) Mul ¶

func (image *Image) Mul(tfout tf.Output) *Image

Mul defines the multiplication operation between the tensor and `tfout`. `tfout` dtype is converted to tensor.Dtype() before multiplying

func (*Image) Normalize ¶

func (image *Image) Normalize() *Image

Normalize computes the mean and the stddev of the pixel values and normalizes every pixel subtracting the mean (centering) and dividing by the stddev (scale)

func (*Image) Pow ¶

func (image *Image) Pow(y tf.Output) *Image

Pow defines the pow operation x^y, where x are the image values y dtype is converted to image.Dtype() before executing Pow

func (*Image) RGBToGrayscale ¶

func (image *Image) RGBToGrayscale() *Image

RGBToGrayscale converts the image from RGB to Grayscale

func (*Image) RGBToHSV ¶

func (image *Image) RGBToHSV() *Image

RGBToHSV performs the colorspace transformation from RGB to HSV

func (*Image) ResizeArea ¶

func (image *Image) ResizeArea(size Size, optional ...op.ResizeAreaAttr) *Image

ResizeArea resizes the image to the specified size using the Area interpolation

func (*Image) ResizeBicubic ¶

func (image *Image) ResizeBicubic(size Size, optional ...op.ResizeBicubicAttr) *Image

ResizeBicubic resizes the image to the specified size using the Bicubic interpolation

func (*Image) ResizeBilinear ¶

func (image *Image) ResizeBilinear(size Size, optional ...op.ResizeBilinearAttr) *Image

ResizeBilinear resizes the image to the specified size using the Bilinear interpolation

func (*Image) ResizeNearestNeighbor ¶

func (image *Image) ResizeNearestNeighbor(size Size, optional ...op.ResizeNearestNeighborAttr) *Image

ResizeNearestNeighbor resizes the image to the specified size using the NN interpolation

func (*Image) SaturateCast ¶

func (image *Image) SaturateCast(dtype tf.DataType) *Image

SaturateCast casts the image to dtype handling overflow and underflow problems, saturate the exceeding values to to minimum/maximum accepted value of the dtype

func (*Image) Scale ¶

func (image *Image) Scale(min, max float32) *Image

Scale scales the image range value to be within [min, max]

func (*Image) Sqrt ¶

func (image *Image) Sqrt() *Image

Sqrt defines the square root operation for the image values

func (*Image) Square ¶

func (image *Image) Square() *Image

Square defines the square operation for the image values

func (*Image) Value ¶

func (image *Image) Value() tf.Output

Value returns the 3D tensor that represents a single image in the Tensorflow environment. If the image is a GIF the returned tensor is 4D

type Point ¶

type Point struct {
	X, Y float32
}

Point represents a single point in the 2D space

type Rectangle ¶

type Rectangle struct {
	Start  Point
	Extent Size
}

Rectangle represents a rectangle in the 2D space. This rectangle starts from the Point Start and has a specified Extent

type Size ¶

type Size struct {
	Height, Width float32
}

Size represents the spatial extent of an Image

type Stride ¶

type Stride struct {
	X, Y int64
}

Stride represents the amount in pixel to move along each dimension

Source Files ¶

View all Source files

Directories ¶

Path	Synopsis
filter
padding

?	: This menu
/	: Search site
f or F	: Jump to
y or Y	: Canonical URL