Um fator importante para o desenvolvimento de projetos de microssistemas é o conhecimento de propriedades termomecânicas dos materiais e a compreensão dos mecanismos de falhas. Este trabalho estuda o comportamento mecânico de microvigas atuadas termicamente e propõem um método para ser utilizado na caracterização de propriedades termomecânicas de filmes finos. Fabricou-se vigas de Oxinitreto de Silício em que se aplicou a microscopia Nomarski para observar a deformação e a ocorrência do fenômeno de flambagem. / An important factor to develop microsystems is knowledge of materials properties and failure mechanisms. This research studies the thermal actuated microbeam mechanical behavior and propose a method in order to characterize thermomechanical properties of thin films. Silicon Oxynitride microbeams are fabricated and Nomarski microscopy was applied to observe strain and buckling phenomenon ocurrence.

Um fator importante para o desenvolvimento de projetos de microssistemas é o conhecimento de propriedades termomecânicas dos materiais e a compreensão dos mecanismos de falhas. Este trabalho estuda o comportamento mecânico de microvigas atuadas termicamente e propõem um método para ser utilizado na caracterização de propriedades termomecânicas de filmes finos. Fabricou-se vigas de Oxinitreto de Silício em que se aplicou a microscopia Nomarski para observar a deformação e a ocorrência do fenômeno de flambagem. / An important factor to develop microsystems is knowledge of materials properties and failure mechanisms. This research studies the thermal actuated microbeam mechanical behavior and propose a method in order to characterize thermomechanical properties of thin films. Silicon Oxynitride microbeams are fabricated and Nomarski microscopy was applied to observe strain and buckling phenomenon ocurrence.

Du fait de leur rapport surface / volume très élevé, la mécanique de objets issus des techniques de micro-fabrication (MEMS) est pilotée par des phénomènes de surface plutôt que de volume. C'est ce qui fait l'intérêt des objets de dimensions micrométriques dans le développement de capteurs environnementaux ou biologiques. <br />L'analyse des phénomènes mis en jeux impose de relever trois défis :<br />Mettre en place une mesure de champs de déplacements nanométriques à l'échelle micrométrique ;<br />Prendre en compte les hétérogénéites des chargements surfaciques et des propriétés mécaniques des objets et valider des modèles de description choisis a priori ;<br />Modéliser de nouveaux couplages, forts, d'origine surfacique.<br />En progressant sur ces trois axes, cette étude a permis de dégager le rôle des principaux paramètres expérimentaux intervenant dans la conception de "puces à ADN" mécaniques.

"Thermal infrared (IR) imaging systems are widely used in medical, industrial, and defense applications. IR imaging systems utilize a lens to focus IR radiation onto a focal plane array (FPA) of IR detectors, which transduce the IR radiation from the scene into signals that can be further processed. In conventional IR imaging systems, electronic readout integrated circuitry (ROIC) is used to read out the information from the FPA, and computer signal processing allows for an IR image to be displayed on an electronic screen. However, the ROIC decreases the thermal isolation and sensitivity of the IR detectors in the FPA, and the computer processing and electronic display increase the cost, weight, and complexity of the IR imaging system. This thesis focuses on the development of an IR direct viewing system that does not require any ROIC, computer signal processing, or electronic display. This is accomplished through the use of microelectromechanical systems (MEMS) uncooled IR imaging detectors, which consist of arrays of bimaterial thermomechanical cantilever structures that tilt as a function of IR radiation from a scene. Other members of the WPI-ME/CHSLT group have previously shown that an interferometric optical readout mechanism based on digital holography and computer processing can eliminate the need for ROIC and be used to measure the nanometer scale tilt of the structures in a MEMS-based IR imaging system that was found to have a responsivity of 1.5 nm/K. However, these previously demonstrated results required significant computer processing and an electronic display. The hypothesis of the current work is that an optomechanical readout mechanism can be used to realize an IR direct viewer without the use of ROIC, computer signal processing, or an electronic display. Three optical readout mechanisms were identified for transducing the nanometer scale deformations of the MEMS structures in the FPA into a directly observable visible light image. Two of these, one using live holography and the other using Nomarski differential interference contrast (DIC), were based on interferometry, while the third, using reflectometry, was based on geometrical optics. The identified optical readout mechanisms were analytically evaluated based on the performance and perception of the human vision system (HVS), and preliminary experimental results were obtained using optical setups constructed for all three readout mechanisms. Based on the analytical and experimental investigations, reflectometry was selected as the most suitable readout mechanism for a direct viewer. A visible light camera was used with custom software to determine a temperature sensitivity of 137 mK for the reflectometry readout, and thermal images of scenes at human body temperature were demonstrated using limited computer processing. A false color, direct view, live IR imaging system was then demonstrated based on a two color reflectometry readout and the output was characterized with respect to the color differentiation sensitivity of the HVS. The system temperature sensitivity, based on the theoretical color differentiation sensitivity of a human observer, was found to be on the order of 10 K across a measuring range of roughly 400 °C, and objects with a temperature as low as approximately 150 °C were distinguishable. The advantages and limitations of the developed IR imaging system are identified and recommendations for further developments and future work are provided."

The ongoing merger of the digital and optical components of the modern microscope is creating opportunities for new measurement techniques, along with new challenges for optical modelling. This thesis investigates several such opportunities and challenges which are particularly relevant to biomedical imaging. Fourier optics is used throughout the thesis as the underlying conceptual model, with a particular emphasis on three--dimensional Fourier optics. A new challenge for optical modelling provided by digital microscopy is the relaxation of traditional symmetry constraints on optical design. An extension of optical transfer function theory to deal with arbitrary lens pupil functions is presented in this thesis. This is used to chart the 3D vectorial structure of the spatial frequency spectrum of the intensity in the focal region of a high aperture lens when illuminated by linearly polarised beam. Wavefront coding has been used successfully in paraxial imaging systems to extend the depth of field. This is achieved by controlling the pupil phase with a cubic phase mask, and thereby balancing optical behaviour with digital processing. In this thesis I present a high aperture vectorial model for focusing with a cubic phase mask, and compare it with results calculated using the paraxial approximation. The effect of a refractive index change is also explored. High aperture measurements of the point spread function are reported, along with experimental confirmation of high aperture extended depth of field imaging of a biological specimen. Differential interference contrast is a popular method for imaging phase changes in otherwise transparent biological specimens. In this thesis I report on a new isotropic algorithm for retrieving the phase from differential interference contrast images of the phase gradient, using phase shifting, two directions of shear, and non--iterative Fourier phase integration incorporating a modified spiral phase transform. This method does not assume that the specimen has a constant amplitude. A simulation is presented which demonstrates good agreement between the retrieved phase and the phase of the simulated object, with excellent immunity to imaging noise.

The ongoing merger of the digital and optical components of the modern microscope is creating opportunities for new measurement techniques, along with new challenges for optical modelling. This thesis investigates several such opportunities and challenges which are particularly relevant to biomedical imaging. Fourier optics is used throughout the thesis as the underlying conceptual model, with a particular emphasis on three--dimensional Fourier optics. A new challenge for optical modelling provided by digital microscopy is the relaxation of traditional symmetry constraints on optical design. An extension of optical transfer function theory to deal with arbitrary lens pupil functions is presented in this thesis. This is used to chart the 3D vectorial structure of the spatial frequency spectrum of the intensity in the focal region of a high aperture lens when illuminated by linearly polarised beam. Wavefront coding has been used successfully in paraxial imaging systems to extend the depth of field. This is achieved by controlling the pupil phase with a cubic phase mask, and thereby balancing optical behaviour with digital processing. In this thesis I present a high aperture vectorial model for focusing with a cubic phase mask, and compare it with results calculated using the paraxial approximation. The effect of a refractive index change is also explored. High aperture measurements of the point spread function are reported, along with experimental confirmation of high aperture extended depth of field imaging of a biological specimen. Differential interference contrast is a popular method for imaging phase changes in otherwise transparent biological specimens. In this thesis I report on a new isotropic algorithm for retrieving the phase from differential interference contrast images of the phase gradient, using phase shifting, two directions of shear, and non--iterative Fourier phase integration incorporating a modified spiral phase transform. This method does not assume that the specimen has a constant amplitude. A simulation is presented which demonstrates good agreement between the retrieved phase and the phase of the simulated object, with excellent immunity to imaging noise.

Η περιοχή Κω – Νισύρου βρίσκεται στο ανατολικότερο άκρο του ενεργού ηφαιστειακού τόξου του νοτίου Αιγαίου, Τεταρτογενούς ηλικίας και περιλαμβάνει τα νησιά Κω και Νίσυρο και τις νησίδες του Γυαλιού, της Στρογγύλης, της Παχειάς και της Πυργούσας. Η ηφαιστειακή δραστηριότητα στη Κω χαρακτηρίζεται από Άνω – Μειοκαινικές ιγκνιμβριτικές (πυρομβριτικές) αποθέσεις και από Πλειοκαινικά – Πλειστοκαινικά ηφαιστειακά προϊόντα όπως τους ηφαιστειακούς δόμους, την πυροκλαστική ακολουθία της Κεφάλου, τους υδρομαγματικούς τόφφους και τον σε μεγάλη έκταση πυροκλαστικό σχηματισμό «τόφφο της Κω» (Kos Plateau Tuff) αποθέσεις του οποίου εντοπίζονται και στα νησιά της Καλύμνου, της Ψερίμου και της Τήλου. Η Νίσυρος είναι ένα στρωματοηφαίστειο δομημένο από Πλειοκαινικά ηφαιστειακά προϊόντα που αποτελούνται κυρίως από ανδεσίτες και βασαλτικούς ανδεσίτες πάνω στους οποίους αποτέθηκαν ασβεσταλκαλικά ηφαιστειακά προϊόντα δακιτικής – ρυοδακιτικής σύστασης με την μορφή πυροκλαστικών αποθέσεων, ροών λάβας και δόμων λάβας. Οι νησίδες του Γυαλιού, της Στρογγύλης, της Παχειάς και της Πυργούσας αποτελούνται από Πλειοκαινικούς δόμους ρυολιθικής (Γυαλί), ανδεσιτικής (Στρογγύλη) και δακιτικής σύστασης (Παχειά και Πυργούσα) οι οποίοι φιλοξενούν πυροκλαστικές αποθέσεις της ανώτερης κίσσηρης του Γυαλιού (Στρογγύλη) και ενότητες του τόφφου της Κω (ΚΡΤ) και της Παναγιάς Κυράς (Παχειά και Πυργούσα). Στη παρούσα διδακτορική διατριβή στόχος είναι η πετρογραφική, ορυκτοχημική και γεωχημική μελέτη των ηφαιστειακών προϊόντων της ευρύτερης περιοχής Κω-Νισύρου και η αποτύπωση ηφαιστειακών δομών και μορφολογικών στοιχείων. Σκοπός είναι η κατανόηση της ηφαιστειακής εξέλιξης με την εξαγωγή συμπερασμάτων που αφορούν τη συγγενετική σχέση των μαγματικών προϊόντων, φαινόμενα μίξης, τους μαγματικούς θαλάμους και την επιφανειακή έκφραση αυτών (καλδέρες). Για το σκοπό αυτό μελετήθηκε η γεωχημική συγγενετική σχέση του τόφφου της Κω (Kos Plateau Tuff) στην Κω – Κάλυμνο – Τήλο και η γεωχημική διαφοροποίηση των πιο βίαιων και εκρηκτικών μονάδων D και E του ιγκνιμβρίτη της Κω επιβεβαιώνοντας τη στρωματογραφική συσχέτιση των αποθέσεων αυτών όπως έχει παρουσιαστεί από δεδομένα κυρίως φυσικής ηφαιστειολογίας. Από την πετροχημική μελέτη της περιοχής Κω – Νισύρου ως ένα ενιαίο «ηφαιστειακό σύμπλεγμα» προκύπτει ότι η ηφαιστειότητα της περιοχής διαχωρίζεται σε Μειοκαινική με προϊόντα όπως ο μονζονίτης και οι Μειοκαινικοί ιγκνιμβρίτες της Κω και σε Πλειο – Πλειοστοκαινική με ηφαιστειακά προϊόντα όπως οι δόμοι και τα πυροκλαστικά από την Κω, το Γυαλί και την Νίσυρο όπου διαφαίνεται η συγγενετική μαγματική σχέση των πετρωμάτων αυτών. Από τη μέθοδο της Συμβολομετρίας «Nomarski» προκύπτει ότι φαινόμενα μαγματικής μίξης/ανάδευσης είναι αναμφισβήτητα και παρέχουν αποδείξεις ότι τα φαινόμενα αυτά ήταν ενεργά καθ’ όλη την διάρκεια της εξέλιξης του ηφαιστείου της Νισύρου. Με την χρήση υπαίθριων παρατηρήσεων και τηλεπισκοπικών μεθόδων εντοπίστηκαν μορφές καλδερών στην ευρύτερη περιοχή Κω – Νισύρου οι οποίες χαρακτηρίσθηκαν και ταξινομήθηκαν σύμφωνα με νέες αντιλήψεις επί της ηφαιστειακής εξέλιξης. Στη συνέχεια μελετήθηκαν οι ηφαιστειακές δομές του ηφαιστείου της Νισύρου με την χρήση της Τηλεπισκόπησης μέσω της φωτοερμηνείας από δορυφορικές εικόνες και ψηφιακά υψομετρικά μοντέλα εδάφους και με την χρήση της γεωμορφομετρικής ανάλυσης συνδυάζοντας γεωμορφο-τεκτονικά χαρακτηριστικά του ηφαιστείου της Νισύρου όπως αυτά παρουσιάζονται από άλλους ερευνητές και νέες αντιλήψεις επί της ηφαιστειακής ανάπτυξη. Επίσης, παρουσιάζεται ένα μοντέλο πρωτο-καλδερικής, καλδερικής και μετα-καλδερικής εξέλιξης του ηφαιστείου της Νισύρου με βάση νέες αντιλήψεις επί της ηφαιστειακής εξέλιξης και σε σχέση με την στρωματογραφική του εξέλιξη. Τέλος, μελετήθηκαν οι υποηφαιστειακοί μαγματικοί θάλαμοι της περιοχής Κω – Νισύρου και τα φαινόμενα μίξης στο ηφαίστειο της Νισύρου με τη μέθοδο της Συμβολομετρίας «Nomarski» και την κατανομή των κύριων στοιχείων, ιχνοστοιχείων και σπανίων γαιών παρουσιάζοντας τον συγγενετικό χαρακτήρα αυτών των κέντρων, ορίζοντας την ευρύτερη ηφαιστειακή περιοχή Κω – Νισύρου ως ένα κύριο ηφαιστειακό κέντρο το οποίο χαρακτηρίζεται από ένα σύνθετο καλδερικό σύστημα. Η αειφορική διαχείριση του ηφαιστείου της Νισύρου είναι το μέσο με το οποίο μπορεί νε συνδυαστεί η ηφαιστειολογική εξέλιξη με την περιβαλλοντική διατήρηση και προβολή της ευρύτερης περιοχής ως γεωτόπου. Από την διερεύνηση της υφιστάμενης κατάστασης του νησιού διαπιστώνεται ότι ενώ η Νίσυρος διαθέτει ένα υψηλό περιβαλλοντικό, ενεργειακό, κοινωνικό, πολιτιστικό και οικονομικό δυναμικό, αυτό παραμένει ανεκμετάλλευτο και αναξιοποίητο λόγω της έλλειψης ενός ολοκληρωμένου σχεδίου αειφορικής διαχείρισης βάση του οποίου θα λαμβάνονται οι αποφάσεις ισότιμα, ισοδύναμα, ταυτόχρονα και αρμονικά. Με βάση την μελέτη των υδατικών πόρων, την καταγραφή του ενεργειακού δυναμικού από ανανεώσιμες πηγές ενέργειας, την διαχείριση του φυσικού και δομημένου περιβάλλοντος του ηφαιστείου της Νισύρου, την μελέτη των επιπτώσεων, θετικών και αρνητικών, της ύπαρξης εξορυκτικής βιομηχανίας στην περιοχή προτάθηκε ένα ολοκληρωμένο σχέδιο αειφόρου ανάπτυξης, προσαρμοσμένο στις απόψεις και τοποθετήσεις, της άμεσα ενδιαφερόμενης, τοπικής κοινωνίας. Από αυτό το ολοκληρωμένο σχέδιο προκύπτουν προτάσεις για την ολοκληρωμένη διαχείριση του ηφαιστείου της Νισύρου οι οποίες ικανοποιούν ισότιμα, ισοδύναμα, ταυτόχρονα και αρμονικά τις αρχές της αειφόρου ανάπτυξης στοχεύοντας στον επαναπροσανατολισμό της τοπικής οικονομίας, την διαχείριση του φυσικού και δομημένου περιβάλλοντος και την κοινωνική ανάπτυξη. / The Kos - Nisyros study area is located at the easternmost edge of the active Quaternary volcanic arc of the southern Aegean Sea and includes the islands of Kos and Nisyros and the islets of Gyali, Strongyli, Pachia and Pyrgoussa. The volcanic activity of Kos island is characterized by Upper – Miocene ignimbrite deposits and Pliocene - Pleistocene volcanic products such as volcanic domes, the Kefalos pyroclastic sequence of hydromagmatic tuffs and the large-scale pyroclastic formation Kos Plateau Tuff (KPT) deposits which is also found on the Kalymnos, Pserimos and Tilos islands. Nisyros is a calcalkaline stratovolcano which consists of Pliocene volcanic products such as andesite and basaltic andesite lavas that are overlain by pyroclastic deposits, lava flows and lava domes of dacitic – rhyodacitic composition. The islets of Gyali, Strongyli, Pachia and Pyrgoussa represent Pliocene lava domes of rhyolitic (Gyali), andesitic (Strogyli) and dacitic composition (Pachia and Pyrgoussa) On these domes, pyroclastic deposits of the Upper Pumice unit of Gyali (in Strogyli) and pyroclastic deposits of the Kos Plateau Tuff (KPT) and Panayia Kyra formation (in Pachia and Pyrgoussa) have been identified. In this thesis, the volcanic products of the Kos-Nisyros area are examined using mineral chemistry, geochemical and petrographical methods, while volcanic structures and morphological features are identified and mapped using remote sensing techniques. The aim of this thesis is to understand the volcanic evolution of the study area, by drawing conclusions regarding the consanguineous relationship of the magmatic products, magma mixing phenomena and the magma chamber system of the area, as well as, the surface expression of this magma chamber system in the form of caldera structures. For this purpose, in this thesis, it is investigated the consanguineous relationship of the Kos Tuff (Kos Plateau Tuff) found on Kos, Tilos and Kalymnos islands, as well as, the geochemical differentiation of the most violent and explosive ignimbrite units D and E of the Kos Tuff, confirming the stratigraphic correlation of these deposits as it was previously presented by physical volcanology data. The petrochemical study of the Kos - Nisyros area as a single "volcanic complex" reveals that the volcanism of the study area is divided into a Miocene and a Pliocene – Pleistocene activity. The Miocene activity is characterized by magmatic and volcanic products such as the monzonite of Kos and the Miocene ignimbrites of Kos and the Pliocene – Pleistocene volcanic activity consists of volcanic products such as lava domes and pyroclastic deposits from Kos, Gyali and Nisyros volcanoes where it is revealed the consanguineous magmatic relationship of these Pliocene – Pleistocene products. Using “Nomarski” interferometry method it is revealed that magma mixing/mingling phenomena are unquestionable for the magma chamber system of Nisyros and provide evidence that these phenomena were active throughout the volcanic evolution of Nisyros volcano. Field observations and remote sensing methods revealed caldera structures in the Kos - Nisyros area which are identified and classified according to new perceptions regarding the volcanic evolution. The volcanic structures of Nisyros volcano are examined using remote sensing methods such as satellite image and digital elevation models interpretation, as well as, the geomorphometric analysis of the Nisyros terrain in combination with geomorphotectonic features of the volcano as presented by other researchers and new perceptions on the volcanic evolution. It is also presented a model of proto-caldera, caldera and post-caldera evolution of Nisyros volcano based on new perceptions on the volcanic evolution in relation to the stratigraphic evolution. Finally, it is studied the subvolcanic magma chamber system of the Kos - Nisyros area, as well as, mixing phenomena in the volcano of Nisyros using the “Nomarski” interferometry method and the distribution of major, trace and rare earth elements revealing the consanguineous nature of these volcanic centers by providing the widest volcanic area of Kos - Nisyros as a major volcanic center characterized by a complex caldera system. Sustainable development of Nisyros volcano represents a management tool which combines the volcanic evolution with the environmental preservation and can promote volcanic regions as geotopes. The investigation of the present conditions of the island reveals, despite the fact that Nisyros is characterized by a strong environmental, energy, social, cultural and economic potential, it remains undeveloped and unexploited due to lack of an integrated sustainable development management plan, based on which, decisions should be equal, equivalent simultaneously and in harmony. Based on the study of the Nisyros water resources, the energy potential from renewable energy sources, the management of natural and urban environment of the volcano of Nisyros and the study of the positive and/or negative effects of the mining industry in Gyali volcanic islet it is proposed an integrated sustainable development according to the views and the perspectives of the local community. From this intergraded plan are generated recommendations for an integrated management of the volcano of Nisyros which satisfy equal, equivalent, simultaneously and in harmony the principles of sustainable development aiming to refocus the local economy, the management of natural and urban environment and the social development.

Dans différents domaines industriels, la problématique du diagnostic prend une place importante. Ainsi, le contrôle d’aspect des composants optiques est une étape incontournable pour garantir leurs performances opérationnelles. La méthode conventionnelle de contrôle par un opérateur humain souffre de limitations importantes qui deviennent insurmontables pour certaines optiques hautes performances. Dans ce contexte, cette thèse traite de la conception d’un système automatique capable d’assurer le contrôle d’aspect. Premièrement, une étude des capteurs pouvant être mis en oeuvre par ce système est menée. Afin de satisfaire à des contraintes de temps de contrôle, la solution proposée utilise deux capteurs travaillant à des échelles différentes. Un de ces capteurs est basé sur la microscopie Nomarski ; nous présentons ce capteur ainsi qu’un ensemble de méthodes de traitement de l’image qui permettent, à partir des données fournies par celui-ci, de détecter les défauts et de déterminer la rugosité, de manière robuste et répétable. L’élaboration d’un prototype opérationnel, capable de contrôler des pièces optiques de taille limitée valide ces différentes techniques. Par ailleurs, le diagnostic des composants optiques nécessite une phase de classification. En effet, si les défauts permanents sont détectés, il en est de même pour de nombreux « faux » défauts (poussières, traces de nettoyage. . . ). Ce problème complexe est traité par un réseau de neurones artificiels de type MLP tirant partie d’une description invariante des défauts. Cette description, issue de la transformée de Fourier-Mellin est d’une dimension élevée qui peut poser des problèmes liés au « fléau de la dimension ». Afin de limiter ces effets néfastes, différentes techniques de réduction de dimension (Self Organizing Map, Curvilinear Component Analysis et Curvilinear Distance Analysis) sont étudiées. On montre d’une part que les techniques CCA et CDA sont plus performantes que SOM en termes de qualité de projection, et d’autre part qu’elles permettent d’utiliser des classifieurs de taille plus modeste, à performances égales. Enfin, un réseau de neurones modulaire utilisant des modèles locaux est proposé. Nous développons une nouvelle approche de décomposition des problèmes de classification, fondée sur le concept de dimension intrinsèque. Les groupes de données de dimensionnalité homogène obtenus ont un sens physique et permettent de réduire considérablement la phase d’apprentissage du classifieur tout en améliorant ses performances en généralisation / In various industrial fields, the problem of diagnosis is of great interest. For example, the check of surface imperfections on an optical device is necessary to guarantee its operational performances. The conventional control method, based on human expert visual inspection, suffers from limitations, which become critical for some high-performances components. In this context, this thesis deals with the design of an automatic system, able to carry out the diagnosis of appearance flaws. To fulfil the time constraints, the suggested solution uses two sensors working on different scales. We present one of them based on Normarski microscopy, and the image processing methods which allow, starting from issued data, to detect the defects and to determine roughness in a reliable way. The development of an operational prototype, able to check small optical components, validates the proposed techniques. The final diagnosis also requires a classification phase. Indeed, if the permanent defects are detected, many “false” defects (dust, cleaning marks. . . ) are emphasized as well. This complex problem is solved by a MLP Artificial Neural Network using an invariant description of the defects. This representation, resulting from the Fourier-Mellin transform, is a high dimensional vector, what implies some problems linked to the “curse of dimensionality”. In order to limit these harmful effects, various dimensionality reduction techniques (Self Organizing Map, Curvilinear Component Analysis and Curvilinear Distance Analysis) are investigated. On one hand we show that CCA and CDA are more powerful than SOM in terms of projection quality. On the other hand, these methods allow using more simple classifiers with equal performances. Finally, a modular neural network, which exploits local models, is developed. We proposed a new classification problems decomposition scheme, based on the intrinsic dimension concept. The obtained data clusters of homogeneous dimensionality have a physical meaning and permit to reduce significantly the training phase of the classifier, while improving its generalization performances