Bootstrap

MLModule

genre	DTIImageProcessing
author	Jan Klein
package	FMEwork/Release
dll	MLDTI
definition	MLDTI.def
see also	AverageSiemensDTI
keywords	DTI, Diffusion, Tensor, bootstrap, bootstrapping

Purpose

The module Bootstrap allows for computing a new artificial DTI data set from a couple of DTI acquisitions (repetitions) -> bootstrap method. It may only be used with Siemens DTI data where gradients are stored in DICOM tags. For computing an averaged data set or trace image from multiple repetitions, please use the related module “AverageSiemensDTI”.

Usage

Connect DTI data to input, select parameters.

• Drawings: how many images should be randomly drawn and then be averaged.

• Output B_0: number preceding B0 images.

• Press button Increase to start a new random selection and a calculation of the output.

Details

Note that you have to use the DecodeVendorSpecificDicomTags module before the Bootstrap module!

Otherwise it will not work. Use DicomTagViewer, enable private tags, and have a look at tag 0021,10xx if the gradient information are stored. B0 images are indicated by \B, gradients by corresponding numbers. Alternatively you may use the DTIInfo module.

Windows

Default Panel

Input Fields

input0

name: input0, type: Image

Siemens DTI data with gradients stored at DICOM tags. At least two repetitions of the measurement have to be stored in the data.

Output Fields

output0

name: output0, type: Image

One single data set with the desired number of preceding B_0 images.

Parameter Fields

Field Index

calculateTrace: Bool

careUnusalDataSorting: Bool

Drawings: Integer

Increase: Trigger

optimalImage: Bool

Output B_0: Integer

Seed value: Integer

Visible Fields

Drawings

name: sampleSize, type: Integer, default: 3, minimum: 1

Sets the number of random drawings for averaging.

Output B_0

name: numberOfB, type: Integer, default: 1

Sets the number of preceding B_0 images of the output data set.

Increase

name: newSeed, type: Trigger

If pressed, a new random selection is started and the output is calculated.

Seed value

name: seedValue, type: Integer, default: 0

Sets a random seed value.

Hidden Fields

optimalImage

name: optimalImage, type: Bool, default: TRUE

If checked, bootstrapping is disabled while averaging is enabled.

calculateTrace

name: calculateTrace, type: Bool, default: FALSE

Disable bootstrapping, compute trace

careUnusalDataSorting

name: careUnusalDataSorting, type: Bool, default: TRUE

If this field is set to True AND if ‘B’ or ‘C’ or ‘ ‘ or ‘' are at the end of the gradient string, then the number of gradients is reduced by one. Usually, this happens only for wrongly implemented DICOM standards. Thus, this value is set to True as default.