mibElasticDistortionFilter.m File Reference

apply elastic deformations to img, the code is based on Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis by Patrice Y. Simard, Dave Steinkraus, John C. Platt http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.160.8494&rep=rep1&type=pdf https://cognitivemedium.com/assets/rmnist/Simard.pdf and codes available at More...

Functions
function [ img , DisplacementField , randomSeed ] =	mibElasticDistortionFilter (img, BatchOpt, randomSeed, DisplacementField)
	apply elastic deformations to img, the code is based on Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis by Patrice Y. Simard, Dave Steinkraus, John C. Platt http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.160.8494&rep=rep1&type=pdf https://cognitivemedium.com/assets/rmnist/Simard.pdf and codes available at

Detailed Description

apply elastic deformations to img, the code is based on Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis by Patrice Y. Simard, Dave Steinkraus, John C. Platt http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.160.8494&rep=rep1&type=pdf https://cognitivemedium.com/assets/rmnist/Simard.pdf and codes available at

• https://stackoverflow.com/questions/39308301/expand-mnist-elastic-deformations-matlab
• https://se.mathworks.com/matlabcentral/fileexchange/66663-elastic-distortion-transformation-on-an-image

Function Documentation

mibElasticDistortionFilter()

function [ img , DisplacementField , randomSeed ] = mibElasticDistortionFilter	(		img,
			BatchOpt,
			randomSeed,
			DisplacementField )

apply elastic deformations to img, the code is based on Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis by Patrice Y. Simard, Dave Steinkraus, John C. Platt http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.160.8494&rep=rep1&type=pdf https://cognitivemedium.com/assets/rmnist/Simard.pdf and codes available at

• https://stackoverflow.com/questions/39308301/expand-mnist-elastic-deformations-matlab
• https://se.mathworks.com/matlabcentral/fileexchange/66663-elastic-distortion-transformation-on-an-image

Parameters

img	image as [height, width, colors, depth]
BatchOpt	structure with parameters .ScalingFactor{1} - [number] defines extension length for deformations .HSize - [char] strel element size for the Gaussian filter .Sigma{1} - [number] sigma value for the Gaussian filter .SourceLayer{1} - [string], type of the dataset, image, model... when not image use nearest interpolation .Mode3D - [logical] distort img in 3D (not yet implemented) .showWaitbar - logical, show or not the waitbar .UseParallelComputing - logical, use or not the parallel computing
randomSeed	[optional], [number], seed to init the random generator
DisplacementField	[optional], a displacement field to use, a structure, should match img [height, width, depth] .fdx .fdy .fdz

Return values

img	image filtered with elastic distortions
DisplacementField	displacement field used in the filter
randomSeed	random seed used to init the field, [] if the field was provided

Required fields of BatchOpt:

Generated fields of DisplacementField:

Referenced by mibDoImageFiltering2().

Here is the caller graph for this function: