|
Aymen El Ghoul
Former PhD Student, UNSA
Keywords : Variational methods, High Order Active Contours, Segmentation, Classification, Object extraction, Line networks, Roads, Rivers, Phase fields Project : Shapes
Contact :
Mail : | | AymendotEl_ghoulatinriadotfr | Phone : | | (33)4-92-38-75-67 | Fax : | | (33)4-92-38-76-43 | Postal adress : | | INRIA Sophia Antipolis
2004, route des Lucioles
06902 Sophia Antipolis Cedex
France | Webpage : | | visit ! |
|
| Abstract :
In my work, I am interested in road network extraction from very high resolution remote sensing images. It has been shown that higher-order active contour (HOAC) models give good results for road extraction from low resolution remotely sensed image. Our aim is to adapt HOAC models to the case of very high resolution remote sensing images which a road can occupy a bidimensional region in image domain. |
Short Bio :
2005 - 2007 : Telecommunications research Master at Ecole Supérieure des Communications de Tunis, Sup'Com (Tunisia)
2003 - 2005 : Telecommunications engineering at Sup'Com
2000 - 2002 : Engineering studies at Institut Préparatoire aux Etudes d'Ingénieur de Tunis, Tunisia. |
Last publications in Ariana Research Group :
A theoretical and numerical study of a phase field higher-order active contour model of directed networks. A. El Ghoul and I. H. Jermyn and J. Zerubia. In The Tenth Asian Conference on Computer Vision (ACCV), Queenstown, New Zealand, November 2010. Keywords : Phase Field, Shape prior, Directed networks, Stability analysis, river extraction, remote sensing. Copyright : Springer-Verlag GmbH Berlin Heidelberg
@INPROCEEDINGS{Elghoul10b,
|
author |
= |
{El Ghoul, A. and Jermyn, I. H. and Zerubia, J.}, |
title |
= |
{A theoretical and numerical study of a phase field higher-order active contour model of directed networks}, |
year |
= |
{2010}, |
month |
= |
{November}, |
booktitle |
= |
{The Tenth Asian Conference on Computer Vision (ACCV)}, |
address |
= |
{Queenstown, New Zealand}, |
pdf |
= |
{http://hal.archives-ouvertes.fr/inria-00522443/fr/}, |
keyword |
= |
{Phase Field, Shape prior, Directed networks, Stability analysis, river extraction, remote sensing} |
} |
Abstract :
We address the problem of quasi-automatic extraction of directed networks, which have characteristic geometric features, from images. To include the necessary prior knowledge about these geometric features, we use a phase field higher-order active contour model of directed networks. The model has a large number of unphysical parameters (weights of energy terms), and can favour different geometric structures for different parameter values. To overcome this problem, we perform a stability analysis of a long, straight bar in order to find parameter ranges that favour networks. The resulting constraints necessary to produce
stable networks eliminate some parameters, replace others by physical parameters such as network branch width, and place lower and upper bounds on the values of the rest.We validate the theoretical analysis via numerical experiments, and then apply the model to the problem of hydrographic network extraction from multi-spectral VHR satellite images. |
Phase fields for network extraction from images. A. El Ghoul. PhD Thesis, Universite de Nice - Sophia-Antipolis, September 2010. Keywords : Shape prior, Higher-order actif contours, Phase Field, Stability analysis, Directed networks, river extraction.
@PHDTHESIS{elghoul10c,
|
author |
= |
{El Ghoul, A.}, |
title |
= |
{Phase fields for network extraction from images}, |
year |
= |
{2010}, |
month |
= |
{September}, |
school |
= |
{Universite de Nice - Sophia-Antipolis}, |
url |
= |
{http://tel.archives-ouvertes.fr/docs/00/55/01/34/PDF/ThesisMunuscript2010_EL_GHOUL.pdf}, |
keyword |
= |
{Shape prior, Higher-order actif contours, Phase Field, Stability analysis, Directed networks, river extraction} |
} |
Résumé :
Cette thèse décrit la construction d'un modèle de réseaux non-directionnels (e.g. réseaux routiers), fondé sur les contours actifs d'ordre supérieur (CAOSs) et les champs de phase développés récemment, et introduit une nouvelle famille des CAOSs des champs de phase pour des réseaux directionnels (e.g. réseaux hydrographiques en imagerie de télédétection, vaisseaux sanguins en imagerie médicale). Dans la première partie de cette thèse, nous nous intéressons à l'analyse de stabilité d'une énergie de type CAOSs aboutissant à un ‘diagramme de phase'. Les résultats, qui sont confirmés par des expériences numériques, permettent une bonne sélection des valeurs des paramètres pour la modélisation de réseaux non-directionnels.
Au contraire des réseaux routiers, les réseaux hydrographiques sont directionnels, i.e. ils contiennent un ‘flux' monodimensionnel circulant dans chaque branche. Cela implique des propriétés géométriques spécifiques des branches et particulièrement des jonctions, propriétés qu'il est utile de traduire dans un modèle, pour l'extraction de réseaux. Nous développons donc un modèle de champ de phase non-local de réseaux directionnels, qui, en plus du champ de phase scalaire décrivant une région par une fonction caractéristique lisse et qui interagit non-localement afin que des configurations de réseaux linéiques soient favorisées, introduit un champ vectoriel représentant le ‘flux' dans les branches du réseau. Ce champ vectoriel est contraint d'être nul à l'extérieur, et de magnitude égale à 1 à l'intérieur du réseau ; circulant dans le sens longitudinal des branches du réseau ; et de divergence très faible. Cela prolonge les branches du réseau ; contrôle la variation de largeur tout au long une branche ; et forme des jonctions non-symétriques telles que la somme des largeurs entrantes soit approximativement égale à celle des largeurs sortantes. En conjonction avec une nouvelle fonction d'interaction pour le champ de phase scalaire, le modèle assure aussi une vaste gamme de valeurs des largeurs stables des branches. Ce nouveau modèle a été appliqué au problème d'extraction de réseaux hydrographiques à partir d'images satellitaires très haute résolution. |
Abstract :
This thesis describes the construction of an undirected network (e.g. road network) model, based on the recently developed higher-order active contours (HOACs) and phase fields, and introduces a new family of phase field HOACs for directed networks (e.g. hydrographic networks in remote sensing imagery, vascular networks in medical imagery). In the first part of this thesis, we focus on the stability analysis of a HOAC energy leading to a ‘phase diagram'. The results, which are confirmed by numerical experiments, enable the selection of parameter values for the modelling of undirected networks.
Hydrographic networks, unlike road networks, are directed, i.e. they carry a unidirectional flow in each branch. This leads to specific geometric properties of the branches and particularly of the junctions, that it is useful to capture in a model, for network extraction purposes. We thus develop a nonlocal phase field model of directed networks, which, in addition to a scalar field representing a region by its smoothed characteristic function, and interacting nonlocally so as to favour network configurations, contains a vector field representing the ‘flow' through the network branches. The vector field is strongly encouraged to be zero outside, and of unit magnitude inside the network; and to have zero divergence. This prolongs network branches; controls width variation along a branch; and produces asymmetric junctions for which total incoming branch width approximately equals total outgoing branch width. In conjunction with a new interaction function for the scalar field, it also allows a broad range of stable branch widths. The new proposed model is applied to the problem of hydrographic network extraction from VHR satellite images, and it outperforms the undirected network model. |
Segmentation of networks from VHR remote sensing images using a directed phase field HOAC model. A. El Ghoul and I. H. Jermyn and J. Zerubia. In Proc. ISPRS Technical Commission III Symposium on Photogrammetry Computer Vision and Image Analysis (PCV), Paris, France, September 2010. Keywords : Phase Field, Shape prior, Directed networks, Road network extraction, river extraction, remote sensing. Copyright : ISPRS
@INPROCEEDINGS{Elghoul10a,
|
author |
= |
{El Ghoul, A. and Jermyn, I. H. and Zerubia, J.}, |
title |
= |
{Segmentation of networks from VHR remote sensing images using a directed phase field HOAC model}, |
year |
= |
{2010}, |
month |
= |
{September}, |
booktitle |
= |
{Proc. ISPRS Technical Commission III Symposium on Photogrammetry Computer Vision and Image Analysis (PCV)}, |
address |
= |
{Paris, France}, |
pdf |
= |
{https://hal.inria.fr/inria-00491017}, |
keyword |
= |
{Phase Field, Shape prior, Directed networks, Road network extraction, river extraction, remote sensing} |
} |
Abstract :
We propose a new algorithm for network segmentation from VHR remote sensing images. The algorithm performs this task quasi-automatically,
that is, with no human intervention except to fix some parameters. The task is made difficult by the amount of prior knowledge about network region geometry needed to perform the task, knowledge that is usually provided by a human being. To include such prior knowledge, we make use of methodological advances in region modelling: a phase field higher-order active contour of directed networks is used as the prior model for region geometry. By adjoining an approximately conserved flow to a phase field model encouraging network shapes (i.e. regions composed of branches meeting at junctions), the model favours network regions in which different branches may have very different widths, but in which width change along a branch is slow; in which branches do not
come to an end, hence tending to close gaps in the network; and in which junctions show approximate ‘conservation of width’. We also introduce image models for network and background, which are validated using maximum likelihood segmentation against other possibilities. We then test the full model on VHR optical and multispectral satellite images. |
|
All publications in Ariana Research Group
|
|