Advanced

Towards Fully Automatic Optimal Shape Modeling

Karlsson, Johan LU (2008)
Abstract
Shape models and the automatic building of such models have proven over

the last decades to be powerful tools in image segmentation and analysis.

This thesis makes contributions to this field.



The segmentation algorithm typically uses an objective function summing

up contributions from each sample point.

In this thesis this is replaced by the approximation of a surface

integral which improves the segmentation results.



Before building a model the shapes in the training set have to be aligned.

This is normally done using Procrustes analysis.

In the thesis an alignment method based on Minimum Decsription Length

(MDL) is examined and... (More)
Shape models and the automatic building of such models have proven over

the last decades to be powerful tools in image segmentation and analysis.

This thesis makes contributions to this field.



The segmentation algorithm typically uses an objective function summing

up contributions from each sample point.

In this thesis this is replaced by the approximation of a surface

integral which improves the segmentation results.



Before building a model the shapes in the training set have to be aligned.

This is normally done using Procrustes analysis.

In the thesis an alignment method based on Minimum Decsription Length

(MDL) is examined and the gradient of MDL is derived and used in the

optimization.



When trying to build optimal models by optimizing MDL there is a

tendency for the parameterizations to put most of their weight on small

parts of the shapes by doing a mutual reparameterization.

In this thesis this problem is solved by replacing the standard scalar

product with a formula that is invariant to mutual reparameterizations.

This is shown to result in better models.



To evaluate the quality of shape models, the standard measures have been

generality, specificity and compactness.

In this thesis, these measures are shown to have severe weaknesses.

An alternative measure called Ground Truth Correspondence Measure is

presented.

This measure is shown to perform better.



Typically, shape modeling assumes that the training set consists of

images where the shape has been segmented as a curve or a surface as a

preprocessing step.

In this thesis a method is introduced that does not need preprocessed

manually segmented data, automatically handles outliers/background and

missing data, and still produces strong models.

The algorithm makes all the decisions about what to include in the model

and what to consider as background and about what points in the

different images are to be considered to be corresponding.

This results in patch-based shape and appearance models generated fully

automatically. (Less)
Please use this url to cite or link to this publication:
author
supervisor
opponent
  • Professor Larsen, Rasmus, Department of Mathematical Modelling, Technical University of Denmark
organization
publishing date
type
Thesis
publication status
published
subject
keywords
Parameterization Invariance, Benchmarking, Interpretation, Segmentation, Alignment, Shape Modeling, MDL
pages
180 pages
defense location
Room MH:C, Centre for Mathematical Sciences, Sölvegatan 18, Lund University Faculty of Engineering
defense date
2008-12-05 13:15
ISBN
978-91-628-7226-7
language
English
LU publication?
yes
id
8c859735-e58b-4c67-9029-9bf44f1a7ad1 (old id 1266177)
date added to LUP
2008-11-10 15:52:13
date last changed
2016-09-19 08:45:17
@misc{8c859735-e58b-4c67-9029-9bf44f1a7ad1,
  abstract     = {Shape models and the automatic building of such models have proven over<br/><br>
the last decades to be powerful tools in image segmentation and analysis.<br/><br>
This thesis makes contributions to this field.<br/><br>
<br/><br>
The segmentation algorithm typically uses an objective function summing<br/><br>
up contributions from each sample point.<br/><br>
In this thesis this is replaced by the approximation of a surface<br/><br>
integral which improves the segmentation results.<br/><br>
<br/><br>
Before building a model the shapes in the training set have to be aligned.<br/><br>
This is normally done using Procrustes analysis.<br/><br>
In the thesis an alignment method based on Minimum Decsription Length<br/><br>
(MDL) is examined and the gradient of MDL is derived and used in the<br/><br>
optimization.<br/><br>
<br/><br>
When trying to build optimal models by optimizing MDL there is a<br/><br>
tendency for the parameterizations to put most of their weight on small<br/><br>
parts of the shapes by doing a mutual reparameterization.<br/><br>
In this thesis this problem is solved by replacing the standard scalar<br/><br>
product with a formula that is invariant to mutual reparameterizations.<br/><br>
This is shown to result in better models.<br/><br>
<br/><br>
To evaluate the quality of shape models, the standard measures have been<br/><br>
generality, specificity and compactness.<br/><br>
In this thesis, these measures are shown to have severe weaknesses.<br/><br>
An alternative measure called Ground Truth Correspondence Measure is<br/><br>
presented.<br/><br>
This measure is shown to perform better.<br/><br>
<br/><br>
Typically, shape modeling assumes that the training set consists of<br/><br>
images where the shape has been segmented as a curve or a surface as a<br/><br>
preprocessing step.<br/><br>
In this thesis a method is introduced that does not need preprocessed<br/><br>
manually segmented data, automatically handles outliers/background and<br/><br>
missing data, and still produces strong models.<br/><br>
The algorithm makes all the decisions about what to include in the model<br/><br>
and what to consider as background and about what points in the<br/><br>
different images are to be considered to be corresponding.<br/><br>
This results in patch-based shape and appearance models generated fully<br/><br>
automatically.},
  author       = {Karlsson, Johan},
  isbn         = {978-91-628-7226-7},
  keyword      = {Parameterization Invariance,Benchmarking,Interpretation,Segmentation,Alignment,Shape Modeling,MDL},
  language     = {eng},
  pages        = {180},
  title        = {Towards Fully Automatic Optimal Shape Modeling},
  year         = {2008},
}