uvaSegTrain: Unsupervised variational autoencoder training

In ANTsX/ANTsRNet: Neural Networks for Medical Image Processing

Unsupervised variational autoencoder training

Description

Trains a variational autoencoding with a convolutional network. This is followed by k-means clustering to produce a segmentation and probabilities.

Usage

uvaSegTrain(patches, k, convControl, standardize = TRUE, patches2)

Arguments

patches	input patch matrix, see getNeighborhoodInMask
k	number of embedding layers
convControl	optional named list with control parameters ( see code ) hiddenAct activation function for hidden layers eg relu img_chns eg 1 number of channels filters eg 32L conv_kern_sz eg 1L front_kernel_size eg 2L intermediate_dim eg 32L epochs eg 50 batch_size eg 32 squashAct activation function for squash layers eg sigmoid tensorboardLogDirectory tensorboard logs stored here
standardize	boolean controlling whether patches are standardized
patches2	input target patch matrix, see getNeighborhoodInMask, may be useful for super-resolution

Value

model is output

Author(s)

Avants BB

Examples

## Not run: 

library(ANTsR)
img <- ri( 1 ) %>% resampleImage( c(4,4) ) %>% iMath( "Normalize" )
mask = randomMask( getMask( img ), 50 )
r = c( 3, 3 )
patch = getNeighborhoodInMask( img, mask, r, boundary.condition = "NA" )
uvaSegModel = uvaSegTrain( patch, 6 )

## End(Not run)

ANTsX/ANTsRNet documentation built on Nov. 21, 2024, 4:07 a.m.