me | department | university | teaching | research | disclaimer

Segmentation

Anatomical model


Anatomical model. This model was used by the MRI Simulator to generate the MRI data sets with different noise levels.

Segmentation results

ML Clusterer





MRI data sets with 3%, 5%, 7% and 9% noise.	Segmented image	Variance image (of the segmented image)	Variance image of the anatomical model

MAP Algorithm





MRI data sets with 3%, 5%, 7% and 9% noise.	Segmented image	Variance image (of the segmented image)	Variance image of the anatomical model

Validation

ML Clusterer


Overall Error Diagram for 3% Noise MRI
x-axis: number of segmented features; y-axis: error rate (and variance from 50 segmentation results).


Overall Error Diagram for 5% Noise MRI
x-axis: number of segmented features; y-axis: error rate (and variance from 50 segmentation results).


Overall Error Diagram for 7% Noise MRI
x-axis: number of segmented features; y-axis: error rate (and variance from 50 segmentation results).


Overall Error Diagram for 9% Noise MRI
x-axis: number of segmented features; y-axis: error rate (and variance from 50 segmentation results).

MAP Algorithm


Mutual Information Diagram for 3% Noise MRI
x-axis: Gibbs potential ß; y-axis: Mutual Information (and variance from 50 segmentation results)


Mutual Information Diagram for 5% Noise MRI
x-axis: Gibbs potential ß; y-axis: Mutual Information (and variance from 50 segmentation results)


Mutual Information Diagram for 7% Noise MRI
x-axis: Gibbs potential ß; y-axis: Mutual Information (and variance from 50 segmentation results)


Mutual Information Diagram for 9% Noise MRI
x-axis: Gibbs potential ß; y-axis: Mutual Information (and variance from 50 segmentation results)

Copyright © 2024 Martin Held, held@cs.sbg.ac.at. All rights reserved.
Impressum and Disclaimer.

file last modified: Thursday, 27-Feb-2020 07:29:07 CET