Advanced Kernel Methods For Remote Sensing Image Classification

Résumé Suite aux recentes avancées technologiques, les archives d'images digitales ont connu une croissance qualitative et quantitative sans précédent. Malgré les énormes possibilités qu'elles offrent, ces avancées posent de nouvelles questions quant au traitement des masses de données saisies. Cette question est à la base de cette Thèse: les problèmes de traitement d'information digitale à très haute résolution spatiale et/ou spectrale y sont considérés en recourant à des approches d'apprentissage statistique, les méthodes à noyau. Cette Thèse étudie des problèmes de classification d'images, c'est à dire de catégorisation de pixels en un nombre réduit de classes refletant les propriétés spectrales et contextuelles des objets qu'elles représentent. L'accent est mis sur l'efficience des algorithmes, ainsi que sur leur simplicité, de manière à augmenter leur potentiel d'implementation pour les utilisateurs. De plus, le défi de cette Thèse est de rester proche des problèmes concrets des utilisateurs d'images satellite sans pour autant perdre de vue l'intéret des méthodes proposées pour le milieu du machine learning dont elles sont issues. En ce sens, ce travail joue la carte de la transdisciplinarité en maintenant un lien fort entre les deux sciences dans tous les développements proposés. Quatre modèles sont proposés: le premier répond au problème de la haute dimensionalité et de la redondance des données par un modèle optimisant les performances en classification en s'adaptant aux particularités de l'image. Ceci est rendu possible par un système de ranking des variables (les bandes) qui est optimisé en même temps que le modèle de base: ce faisant, seules les variables importantes pour résoudre le problème sont utilisées par le classifieur. Le manque d'information étiquétée et l'incertitude quant à sa pertinence pour le problème sont à la source des deux modèles suivants, basés respectivement sur l'apprentissage actif et les méthodes semi-supervisées: le premier permet d'améliorer la qualité d'un ensemble d'entraînement par interaction directe entre l'utilisateur et la machine, alors que le deuxième utilise les pixels non étiquetés pour améliorer la description des données disponibles et la robustesse du modèle. Enfin, le dernier modèle proposé considère la question plus théorique de la structure entre les outputs: l'intègration de cette source d'information, jusqu'à présent jamais considérée en télédétection, ouvre des nouveaux défis de recherche. Advanced kernel methods for remote sensing image classification Devis Tuia Institut de Géomatique et d'Analyse du Risque September 2009 Abstract The technical developments in recent years have brought the quantity and quality of digital information to an unprecedented level, as enormous archives of satellite images are available to the users. However, even if these advances open more and more possibilities in the use of digital imagery, they also rise several problems of storage and treatment. The latter is considered in this Thesis: the processing of very high spatial and spectral resolution images is treated with approaches based on data-driven algorithms relying on kernel methods. In particular, the problem of image classification, i.e. the categorization of the image's pixels into a reduced number of classes reflecting spectral and contextual properties, is studied through the different models presented. The accent is put on algorithmic efficiency and the simplicity of the approaches proposed, to avoid too complex models that would not be used by users. The major challenge of the Thesis is to remain close to concrete remote sensing problems, without losing the methodological interest from the machine learning viewpoint: in this sense, this work aims at building a bridge between the machine learning and remote sensing communities and all the models proposed have been developed keeping in mind the need for such a synergy. Four models are proposed: first, an adaptive model learning the relevant image features has been proposed to solve the problem of high dimensionality and collinearity of the image features. This model provides automatically an accurate classifier and a ranking of the relevance of the single features. The scarcity and unreliability of labeled. information were the common root of the second and third models proposed: when confronted to such problems, the user can either construct the labeled set iteratively by direct interaction with the machine or use the unlabeled data to increase robustness and quality of the description of data. Both solutions have been explored resulting into two methodological contributions, based respectively on active learning and semisupervised learning. Finally, the more theoretical issue of structured outputs has been considered in the last model, which, by integrating outputs similarity into a model, opens new challenges and opportunities for remote sensing image processing.

ZHARP: three-dimensional motion tracking from a single image plane.

Three-dimensional imaging and quantification of myocardial function are essential steps in the evaluation of cardiac disease. We propose a tagged magnetic resonance imaging methodology called zHARP that encodes and automatically tracks myocardial displacement in three dimensions. Unlike other motion encoding techniques, zHARP encodes both in-plane and through-plane motion in a single image plane without affecting the acquisition speed. Postprocessing unravels this encoding in order to directly track the 3-D displacement of every point within the image plane throughout an entire image sequence. Experimental results include a phantom validation experiment, which compares zHARP to phase contrast imaging, and an in vivo study of a normal human volunteer. Results demonstrate that the simultaneous extraction of in-plane and through-plane displacements from tagged images is feasible.

Partition coefficient and molecular flexibility: the concept of lipophilicity space.

The rationale of this study was to investigate molecular flexibility and its influence on physicochemical properties with a view to uncovering additional information on the fuzzy concept of dynamic molecular structure. Indeed, it is now known that computed molecular interaction fields (MIFs) such as molecular electrostatic potentials (MEPs) and lipophilicity potentials (MLPs) are conformation-dependent, as are dipole moments. A database of 125 compounds was used whose conformational space was explored, while conformation-dependent parameters were computed for each non-redundant conformer found in the conformational space of the compounds. These parameters were the virtual log P (log P(MLP), calculated by a MLP approach), the apolar surface area (ASA), polar surface area (PSA), and solvent-accessible surface (SAS). For each compound, the range taken by each parameter (its property space) was divided by the number of rotors taken as an index of flexibility, yielding a parameter termed 'molecular sensitivity'. This parameter was poorly correlated with others (i.e., it contains novel information) and showed the compounds to fall into two broad classes. 'Sensitive' molecules are those whose computed property ranges are markedly sensitive to conformational effects, whereas 'insensitive' (in fact, less sensitive) molecules have property ranges which are comparatively less affected by conformational fluctuations. A pharmacokinetic application is presented.

Postmortem dynamic CT-angiography : P30

Purpose: Forensic imaging and especially forensic radiology is a new trend in forensic medicine. More and more forensic institutes set up their own CT-scanner in order to perform postmortem cross-sectional imaging. Due to this trend, a new subspecialty was born: the forensic radiology. To image the vascular system after death, a postmortem CT- angiography can be performed. Methods and materials: In the Institute of Forensic Medicine in Lausanne, a science group has been created with specialists of different medical fields that has set up a new technique of forensic CT-angiography. The method consists in the creation of a postmortem circulation by the use of a modified heart lung machine. As circulating liquid Angiofil, an oily contrast agent, is injected. Results: With the aid of this technique, the whole vascular system of a deceased person can be imaged in detail without autopsy. The circulating contrast allows demonstrating the vascular system when it is under pressure, similarly to living patients. First experiences showed, that vascular pathologies such as cardiac tamponade and aortic dissection can be well demonstrated. Since the oily Angiofil strictly remains in the intravascular space, no artifacts had been observed during the CT-examination and the later performed autopsy. Conclusion: Post-mortem dynamic CT angiography is of great advantage in forensic pathology, because the detailed mapping of the entire vascular system is almost impossible with conventional autopsy tools.

Shift in cytotype frequency and niche space in the invasive plant Centaurea maculosa.

Polyploidy is often assumed to increase the spread and thus the success of alien plant species, but few empirical studies exist. We tested this hypothesis with Centaurea maculosa Lam., a species native to Europe and introduced into North America approximately 120 years ago where it became highly invasive. We analyzed the ploidy level of more than 2000 plants from 93 native and 48 invasive C. maculosa populations and found a pronounced shift in the relative frequency of diploid and tetraploid cytotypes. In Europe diploid populations occur in higher frequencies than tetraploids and only four populations had both cytotypes, while in North America diploid plants were found in only one mixed population and thus tetraploids clearly dominated. Our results showed a pronounced shift in the climatic niche between tetraploid populations in the native and introduced range toward drier climate in North America and a similar albeit smaller shift between diploids and tetraploids in the native range. The field data indicate that diploids have a predominately monocarpic life cycle, while tetraploids are often polycarpic. Additionally, the polycarpic life-form seems to be more prevalent among tetraploids in the introduced range than among tetraploids in the native range. Our study suggests that both ploidy types of C. maculosa were introduced into North America, but tetraploids became the dominant cytotype with invasion. We suggest that the invasive success of C. maculosa is partly due to preadaptation of the tetraploid cytotype in Europe to drier climate and possibly further adaptation to these conditions in the introduced range. The potential for earlier and longer seed production associated with the polycarpic life cycle constitutes an additional factor that may have led to the dominance of tetraploids over diploids in the introduced range.