title | description | author | ms.author | ms.date | ms.service | ms.subservice | ms.topic | keywords | monikerRange | ||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
extractPixels function (MicrosoftML) |
Extracts the pixel values from an image (MicrosoftML). |
rothja |
jroth |
07/15/2019 |
sql |
machine-learning |
reference |
|
>=sql-server-2016||>=sql-server-linux-ver15 |
Extracts the pixel values from an image.
extractPixels(vars, useAlpha = FALSE, useRed = TRUE, useGreen = TRUE,
useBlue = TRUE, interleaveARGB = FALSE, convert = TRUE, offset = NULL,
scale = NULL)
A named list of character vectors of input variable names and the name of the output variable. Note that the input variables must be of the same type. For one-to-one mappings between input and output variables, a named character vector can be used.
Specifies whether to use alpha channel. The default value is FALSE
.
Specifies whether to use red channel. The default value is TRUE
.
Specifies whether to use green channel. The default value is TRUE
.
Specifies whether to use blue channel. The default value is TRUE
.
Whether to separate each channel or interleave in ARGB order. This might be important, for example, if you are training a convolutional neural network, since this would affect the shape of the kernel, stride etc.
Whether to convert to floating point. The default value is FALSE
.
Specifies the offset (pre-scale). This requires convert = TRUE
. The default value is NULL
.
Specifies the scale factor. This requires convert = TRUE
. The default value is NULL
.
extractPixels
extracts the pixel values from an image. The input variables
are images of the same size, typically the output of a resizeImage
transform. The
output is pixel data in vector form that are typically used as features for a learner.
A maml
object defining the transform.
Microsoft Corporation Microsoft Technical Support
train <- data.frame(Path = c(system.file("help/figures/RevolutionAnalyticslogo.png", package = "MicrosoftML")), Label = c(TRUE), stringsAsFactors = FALSE)
# Loads the images from variable Path, resizes the images to 1x1 pixels and trains a neural net.
model <- rxNeuralNet(
Label ~ Features,
data = train,
mlTransforms = list(
loadImage(vars = list(Features = "Path")),
resizeImage(vars = "Features", width = 1, height = 1, resizing = "Aniso"),
extractPixels(vars = "Features")
),
mlTransformVars = "Path",
numHiddenNodes = 1,
numIterations = 1)
# Featurizes the images from variable Path using the default model, and trains a linear model on the result.
model <- rxFastLinear(
Label ~ Features,
data = train,
mlTransforms = list(
loadImage(vars = list(Features = "Path")),
resizeImage(vars = "Features", width = 224, height = 224), # If dnnModel == "AlexNet", the image has to be resized to 227x227.
extractPixels(vars = "Features"),
featurizeImage(var = "Features")
),
mlTransformVars = "Path")