2-D implementation of the VGG deep learning architecture.

Creates a keras model of the Vgg deep learning architecture for image recognition based on the paper

createVggModel2D(
  inputImageSize,
  numberOfClassificationLabels = 1000,
  layers = c(1, 2, 3, 4, 4),
  lowestResolution = 64,
  convolutionKernelSize = c(3, 3),
  poolSize = c(2, 2),
  strides = c(2, 2),
  numberOfDenseUnits = 4096,
  dropoutRate = 0,
  style = 19,
  mode = c("classification", "regression")
)

Arguments

inputImageSize	Used for specifying the input tensor shape. The shape (or dimension) of that tensor is the image dimensions followed by the number of channels (e.g., red, green, and blue). The batch size (i.e., number of training images) is not specified a priori.
numberOfClassificationLabels	Number of segmentation labels.
layers	a vector determining the number of filters defined at for each layer.
lowestResolution	number of filters at the beginning.
convolutionKernelSize	2-d vector definining the kernel size during the encoding path
poolSize	2-d vector defining the region for each pooling layer.
strides	2-d vector describing the stride length in each direction.
numberOfDenseUnits	integer for the number of units in the last layers.
dropoutRate	float between 0 and 1 to use between dense layers.
style	'16' or '19' for VGG16 or VGG19, respectively.
mode	'classification' or 'regression'.

Value

a VGG keras model

Details

K. Simonyan and A. Zisserman, Very Deep Convolutional Networks for Large-Scale Image Recognition

available here:

    \url{https://arxiv.org/abs/1409.1556}

This particular implementation was influenced by the following python implementation:

    \url{https://gist.github.com/baraldilorenzo/8d096f48a1be4a2d660d}

Author

Tustison NJ

Examples

library( ANTsRNet )
library( keras )
library( ANTsR )

mnistData <- dataset_mnist()
#> Error in py_discover_config(required_module, use_environment): Python specified in RETICULATE_PYTHON (/Users/ntustison/anaconda3/envs/antsx/bin/python3) does not exist
numberOfLabels <- 10

# Extract a small subset for something that can run quickly.
# We also need to resample since the native mnist data size does
# not fit with GoogLeNet parameters.

resampledImageSize <- c( 100, 100 )
numberOfTrainingData <- 10
numberOfTestingData <- 5

X_trainSmall <- as.array(
  resampleImage( as.antsImage( mnistData$train$x[1:numberOfTrainingData,,] ),
    c( numberOfTrainingData, resampledImageSize ), TRUE ) )
#> Error in h(simpleError(msg, call)): error in evaluating the argument 'x' in selecting a method for function 'check_ants': error in evaluating the argument 'object' in selecting a method for function 'as.antsImage': object 'mnistData' not found
X_trainSmall <- array( data = X_trainSmall, dim = c( dim( X_trainSmall ), 1 ) )
#> Error in array(data = X_trainSmall, dim = c(dim(X_trainSmall), 1)): object 'X_trainSmall' not found
Y_trainSmall <- to_categorical( mnistData$train$y[1:numberOfTrainingData], numberOfLabels )
#> Error in to_categorical(mnistData$train$y[1:numberOfTrainingData], numberOfLabels): object 'mnistData' not found

X_testSmall <- as.array(
  resampleImage( as.antsImage( mnistData$test$x[1:numberOfTestingData,,] ),
    c( numberOfTestingData, resampledImageSize ), TRUE ) )
#> Error in h(simpleError(msg, call)): error in evaluating the argument 'x' in selecting a method for function 'check_ants': error in evaluating the argument 'object' in selecting a method for function 'as.antsImage': object 'mnistData' not found
X_testSmall <- array( data = X_testSmall, dim = c( dim( X_testSmall ), 1 ) )
#> Error in array(data = X_testSmall, dim = c(dim(X_testSmall), 1)): object 'X_testSmall' not found
Y_testSmall <- to_categorical( mnistData$test$y[1:numberOfTestingData], numberOfLabels )
#> Error in to_categorical(mnistData$test$y[1:numberOfTestingData], numberOfLabels): object 'mnistData' not found

# We add a dimension of 1 to specify the channel size

inputImageSize <- c( dim( X_trainSmall )[2:3], 1 )
#> Error in eval(expr, envir, enclos): object 'X_trainSmall' not found

model <- createVggModel2D( inputImageSize = c( resampledImageSize, 1 ),
  numberOfClassificationLabels = numberOfLabels )
#> Error in py_discover_config(required_module, use_environment): Python specified in RETICULATE_PYTHON (/Users/ntustison/anaconda3/envs/antsx/bin/python3) does not exist

model %>% compile( loss = 'categorical_crossentropy',
  optimizer = optimizer_adam( lr = 0.0001 ),
  metrics = c( 'categorical_crossentropy', 'accuracy' ) )
#> Error in compile(., loss = "categorical_crossentropy", optimizer = optimizer_adam(lr = 1e-04),     metrics = c("categorical_crossentropy", "accuracy")): object 'model' not found

# Comment out the rest due to travis build constraints

# track <- model %>% fit( X_trainSmall, Y_trainSmall, verbose = 1,
#   epochs = 1, batch_size = 2, shuffle = TRUE, validation_split = 0.5 )

# Now test the model

# testingMetrics <- model %>% evaluate( X_testSmall, Y_testSmall )
# predictedData <- model %>% predict( X_testSmall, verbose = 1 )