Éditeur(s) : HAL CCSD
Résumé : International audience
This paper presents the early work, done in the context of the IntuiScript project, on handwriting quality analysis. This IntuiScript project aims at developing a digital workbook to help with teaching children how to handwrite. To do so, we must be able to analyse their handwriting, to evaluate if the letters are correctly written, and to detail what aspects of the child symbols – letters, numbers, and geometric forms-do not correspond to the teacher models. We use an online fuzzy model to easily build target models, and to automatically evaluate the adequacy of children letters to these reference models, with respect to different aspects: symbol shape, drawing direction and stroke order for example.
17th Biennial Conference of the International Graphonomics Society
Pointe-à-Pitre, Guadeloupe

Droits : info:eu-repo/semantics/OpenAccess
hal-01165765
https://hal.univ-antilles.fr/hal-01165765
https://hal.univ-antilles.fr/hal-01165765/document
https://hal.univ-antilles.fr/hal-01165765/file/IGS_2015_submission_4.pdf

Éditeur(s) : Université des Antilles Études caribéennes
Résumé : El Caribe colombiano ha experimentado un acelerado proceso de urbanización y desarrollo turístico en los últimos 50 años. Solo en una ciudad como Cartagena de Indias (la ciudad turística por excelencia de Colombia), datos referidos al periodo noviembre 2009 – enero 2015, mostraron un incremento en la llegada de turistas cercano al 20%. Después de un gran número de encuestas realizadas alrededor del mundo, se encontró que los parámetros en que se basan los turistas a la hora de preferir una playa son: i) la seguridad, ii) el paisaje, iii) calidad del agua, iv) basuras, v) instalaciones. El paisaje es uno de los recursos costeros y por lo tanto su evaluación es una herramienta importante para los gestores y planificadores del litoral a la hora de la realización de planes de conservación, protección, desarrollo y manejo. Este trabajo se basa en la evaluación del valor paisajístico de las playas que constituyen la base de la economía de muchas ciudades costeras en el mundo y en especial de la zona central de litoral Caribe colombiano. Un problema importante en la evaluación del paisaje es la incapacidad de algunas metodologías de reflejar correctamente las percepciones de la gente, debido al alto peso que tienen algunos aspectos subjetivos (p. ej. altura de una geoforma).En este trabajo se presenta la evaluación del paisaje realizada para la zona central de Caribe colombiano (37 playas de los departamentos de Sucre, Bolívar y Atlántico) utilizando la lógica matemática y matrices ponderadas (fuzzy logic mathematics - parameter weighting matrices) las cuales permiten disminuir la subjetividad y cuantificar con exactitud la incertidumbre. Para la evaluación se utilizó una lista de chequeo que está conformada por 26 parámetros, 18 físicos (acantilados, playas, plataformas rocosas, dunas, características de los valles, forma del relieve, mareas, paisaje costero, color del agua y restos de vegetación), y 8 antropogénicos (ruido, basura, descarga de aguas residuales, grado de modificación y de construcción del medio, tipos de acceso, línea del horizonte y estructuras antrópicas). Los parámetros fueron calificados en una escala de 1 a 5 (presencia/ausencia o poca/alta calidad) y, con el fin de cuantificar la incertidumbre y excesos derivados de la subjetividad que pudo haber tenido la evaluación de los parámetros, se utilizó la lógica difusa (fuzzy logic). A la par, un sistema de matrices, se empleó para asignar valores de peso de acuerdo a las preferencias y prioridades de los usuarios. Como resultado de estos análisis lógico-estadísticos se obtuvo un valor (D) el cual resume la evaluación del paisaje en 5 clases que van desde la CLASE 1 (zona litoral sumamente atractiva) hasta la CLASE 5 (playas urbanas muy poco atractivas). De los 37 sectores estudiados, ninguno puede ser considerado como una zona litoral sumamente atractiva, mientras que 15 corresponden a zonas litorales urbanas poco atractivas, el resto de las playas (22) se ubicó entre las clases 2, 3 y 4. Los resultados obtenidos son útiles para los gestores costeros en la medida que, aunque no es posible mejorar los aspectos naturales de la clasificación, si es posible mejorar muchos de los parámetros antropogenicos y, de esta forma, hacer que una playa mejore su clasificación y sea paisajísticamente más atractiva.
In the past few years, the Colombian Caribbean coast has experienced rapid tourism development which has not been accompanied with a similar increase in urbanization facilities. For example, the city of Cartagena de Indias (the most famous tourist destination of Colombia), data related to November 2009 – January 2015 period, showed an increase of 20% in tourist arrivals. A large number of surveys developed in Europe found that the most critical parameters for tourists, when they choose a beach, are: i) security, ii) scenery, iii) water quality, iv) litter, v) facilities. Scenery is one of the most significant coastal resources. Therefore, its evaluation is a valuable tool for coastal managers when developing conservation and protection plans. This proposal provides coastal scenery assessment of 37 beaches along the North Colombian Caribbean littoral by analysing 26 physical and human factors. Results were weighted (not all parameters have equal weight) and subjected to fuzzy logic mathematics in order to reduce recorder judgment subjectivity. The sites were categorised into five classes; Class 1 locations had top grade scenery; Class 5 poor scenery. Of the 37 sectors studied, currently none can be considered as an extremely attractive coastal area (class1), while 15 were unattractive urban coastal areas (class 5), the rest of the beaches (22) placed in classes 2, 3 and 4. Classification of analysed sites depends on the geological setting and degree of human occupation. The technique opens new perspectives for analysis of the potential for coastal tourism development in natural areas and scenic quality improvement of current tourist developed areas. Concerning coastal management issues, emphasis should be given to the upgrading of human parameters eliminating litter and sewage evidence, vegetation debris and enhancing beach nourishment works. This proposal can help in moving some locations from their present status to a higher one.
Caraïbes
Colombie

Droits : info:eu-repo/semantics/openAccess
urn:doi:10.4000/etudescaribeennes.9326
http://journals.openedition.org/etudescaribeennes/9326

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
Forecasting of sea-state characteristics, with warning time of a few hours, appears a necessity in Operational Oceanography. linking sophisticated marine monitoring systems with forecasting modeling tools. In this paper, instead of using conventional numerical models, a Takagi-Sugeno-rule-based Fuzzy Inference System (FIS) was developed aiming at forecasting wave parameters based on the wind speed and direction, and the lagged-wave characteristics. Initial and final antecedent fuzzy membership functions were identified using the subtractive clustering method. The model was applied on the wind and wave dataset recorded in years 2000-2006 by an oceanographic buoy deployed in the Aegean Sea. The model showed perfect fit for the training period (2000-2005; 12,274 data points), and expanded its hindcasting ability during 2006 (1044 data points), as the verification part of the series. Model results, for a lead time of 3 h, showed good agreement between the predicted and the observed significant wave height (RMSE=0.216) and zero-up-crossing period (RMSE=0.315). According to other model performance criteria, the fuzzy model slightly underpredicted both wave characteristics (the linear regression slope was 0.911 for wave height and 0.788 for wave period), and reduced its forecasting ability at higher prediction intervals (+6 to + 12 h). Overall, model results illustrated that the developed FIS could serve as a valuable tool for the operational prediction of wave characteristics in Northern Aegean Sea, through the utilization of the POSEIDON network.
ISSN: 0029-8018

hal-00456020
https://hal.archives-ouvertes.fr/hal-00456020
DOI : 10.1016/j.oceaneng.2009.08.016

Éditeur(s) : HAL CCSD
Résumé : International audience
In this paper, we present a novel method for recognition of handwritten sketches. Unlike previous approaches, we focus on online retrieval and ability to build our model incrementally, thus we do not need to know all the data in advance and we can achieve very good recognition results after as few as 15 samples. The method is composed of two main parts: feature representation and learning and recognition. In feature representation part, we utilize SIFT-like feature descriptors in combination with soft response Bag-of-Words techniques. Descriptors are extracted locally using our novel sketch-specific sampling strategy and for support regions we follow patch-based approach. For learning and recognition, we use a novel technique based on fuzzy-neural networks, which has shown good performance in incremental learning. The experiments on state-of-the-art benchmarks have shown promising results.
17th Biennial Conference of the International Graphonomics Society
Pointe-à-Pitre, Guadeloupe

hal-01166497
https://hal.univ-antilles.fr/hal-01166497
https://hal.univ-antilles.fr/hal-01166497/document
https://hal.univ-antilles.fr/hal-01166497/file/IGS_2015_submission_37.pdf

Éditeur(s) : HAL CCSD
Résumé : International audience
This paper deals with data structures within GIS. Continuous phenomenons are usually represented by raster structures for simplicity reasons. With such structures spatial repartitions of the data is not easily interpretable. Moreover, in an overlapping clustering context these structures remove the links between the data and the algorithms. We propose a vector representation of such data based on non-regular multi-rings polygons. The structure requires multipart nested polygons and new set operations. We present the formalism based on belief theory and uncertainty reasoning. We also detail the implementation of the structures and the set operations. The structures and the set operations are illustrated in the context of forest classification having diffuse transitions.
IF&GIS
Shangai, China

hal-01514239
https://hal.archives-ouvertes.fr/hal-01514239

Éditeur(s) : Ecole Nationale Supérieure Agronomique de Rennes
Résumé : Long-line fishing has been growing on the island of La Réunion since the beginning of the 1990s. The target species is the swordfish (Xiphias gladius), fished mainly in the tropical waters surrounding the island between longitudes 45 degree E and 60 degree E, and latitudes 17 degree S and 27 degree S. From 1998 to 2000, a finalised research programme enabled the collection of precise data concerning the lines' positions and the parameters of the fishing efforts: 3,602 nettings were thus informed by the catches (in numbers of individuals caught) and 12 descriptive variables of the effort. In parallel with this collection effort, satellite maps that described the ocean's landscape of this region of the Indian Ocean exploited by La Réunion's fisheries were made available to sailors and scientists. Extractions from these environmental variables were made at the fishing lines' positions, which allowed us to propose 20 variables describing the environmental conditions of swordfish catches applied to the nettings. In an early part of this study, after a descriptive analysis and a first selection of the variables, we used several GAM (Generalised Additive Models) to model the respective effects of effort and environmental variables on the variability of swordfish catches. We thus highlighted the predominance of factors linked to fishermen (fishing effort) on oceanographic factors in the explanation of this variability. Following these results, a summary on the biology and ecology of the swordfish was proposed in order to draw up hypotheses on the characteristics of this species' population in the Indian Ocean. The aspects linked to the individual behaviour and movements of the population were particularly highlighted. In a latter part of this study and in order to complete the empirical approach by an exploratory constructivist approach, we proposed a modelling tool in order to simulate the possible movements of individuals within the oceanic landscape described by the available satellite maps. After reading in the literature about the various forms of modelling the movements of the large pelagics, we developed an original computer modelling architecture of the multi-agent type that made it possible to manipulate in a dynamic way the satellite maps (and derived products), the individual movements of "animats" whose behaviours were based on this environmental information, and agents that clarify the process of long-line fishing in a geographical way. The computer model that was implemented (MUFINS, for MUlti-Fish INdian ocean Simulator) was next used by calibrating the behaviour of the animats on the real data originally from the analysis of commercial catches proposed in the first part, using robust techniques based on fuzzy logic. The results of the simulations were compared, on the one hand, with the large-scale hypotheses of movement proposed in the first part, and, on the other hand, to swordfish catches both real and corrected for the effect of the fishing effort modelled by the GAM models. Even if the results of these simulations in this study do not allow for proposing convincing interpretations of the underlying processes, the MUFINS tool showed its ability to address questions associated with the swordfish's behaviour in the tropical oceanic landscape and to the effects of scale transfers between the individual level and the collective level. We now envisage its use in a more operational framework, as an interface for dialogue between fishermen and scientists with a goal of sustainable management of the resource.
La peche palangriere s'est developpee a l'ile de La Reunion depuis le debut des annees 1990. L'espece cible est l'espadon (Xiphias gladius), principalement exploite dans les eaux tropicales autour de l'ile entre les longitudes 45 degree E et 60 degree E, et les latitudes 17 degree S et 27 degree S. De 1998 a 2000, un programme de recherche finalisee a permis de collecter des informations precises concernant les positions des lignes et les parametres de l'effort de peche : 3602 filages sont ainsi renseignes par les captures (en nombre d'individus captures) et 12 variables descriptives de l'effort. En parallele de cet effort de collecte, des cartes satellitales ont pu etre mises a la disposition des marins et des scientifiques, decrivant le paysage oceanique de cette region de l'ocean Indien exploitee par la pecherie reunionnaise. Des extractions de ces variables environnementales ont ete effectuees aux positions des lignes de peche, nous permettant de proposer 20 variables descriptives des conditions environnementales de captures d'espadon appliquees aux filages. Dans une premiere partie de ce travail, apres une analyse descriptive et une premiere selection des variables, on a modelise par plusieurs modeles GAM (Generalized Additive Models) les effets respectifs des variables d'effort et des variables environnementales sur la variabilite des captures d'espadon. On a ainsi mis en evidence la predominance des facteurs lies aux pecheurs (effort de peche) sur les facteurs oceanographiques dans le caractere explicatif de cette variabilite. A la suite de ces resultats, une synthese sur la biologie et l'ecologie de l'espadon a ete proposee afin de degager des hypotheses sur les caracteristiques de la population de cette espece dans l'ocean Indien. Les aspects lies au comportement individuel et aux deplacements de la population ont ete particulierement mis en evidence. Dans une deuxieme partie de ce travail et afin de completer l'approche empirique par une approche constructiviste exploratoire, nous avons propose un outil de modelisation afin de simuler les deplacements possibles d'individus au sein du paysage oceanique decrit par les cartes satellitales disponibles. Apres avoir passe en revue dans la litterature les differentes formes de modelisation des deplacements des grands, pelagiques, nous avons developpe une architecture originale de modelisation informatique de type multi-agents permettant de manipuler de maniere dynamique les cartes satellitales (et les produits derives), les mouvements individuels < d'animats >, dont les comportements sont bases sur cette information environnementale, et des agents qui explicitent le processus de capture palangriere de maniere geographique. Le modele informatique ainsi implemente (MUFINS, pour MUlti-Fish INdian ocean Simulator) a ensuite ete utilise en calibrant les comportements des animats sur les donnees reelles issues de l'analyse des captures commerciales proposee en premiere partie, a partir de techniques robustes basee sur la logique floue. Les resultats des simulations ont ete confrontes d'une part aux hypotheses de deplacement a grande echelle proposees en premiere partie, et d'autre part aux captures d'espadon reelles et corrigees de l'effet de l'effort de peche modelise par les modeles GAM. Meme si les resultats de simulations presentes dans ce travail ne permettent pas de proposer d'interpretations convaincantes des processus sous-jacents, l'outil MUFINS a demontre son aptitude a aborder les questions liees au comportement de l'espadon dans le paysage oceanique tropical et aux effets des transferts d'echelle entre le niveau individuel et le niveau collectif. On envisage desormais son utilisation dans un cadre plus operationnel, comme interface de dialogue entre les pecheurs et les scientifiques a des fins de gestion durable de la ressource.

Droits : info:eu-repo/semantics/openAccess
http://archimer.ifremer.fr/doc/2005/these-2592.pdf
http://archimer.ifremer.fr/doc/00000/2592/

Éditeur(s) : HAL CCSD Wiley-Blackwell
Résumé : International audience
1. Our understanding of the contribution of interspecific interactions to functional diversity in nature lags behind our knowledge of spatial and temporal patterns. Although two-species mutualisms are found in all types of ecosystems, the study of their ecological influences on other community members has mostly been limited to third species, while their influence on entire communities remains largely unexplored. 2. We hypothesized that mutualistic interactions between two respective ant species and an epiphyte mediate the biological traits composition of entire invertebrate communities that use the same host plant, thereby affecting food webs and functional diversity at the community level. 3. Aechmea mertensii (Bromeliaceae) is both a phytotelm ('plant-held water') and an ant-garden epiphyte. We sampled 111 bromeliads (111 aquatic invertebrate communities) associated with either the ant Pachycondyla goeldii or Camponotus femoratus. The relationships between ants, bromeliads and invertebrate abundance data were examined using a redundancy analysis. Biological traits information for invertebrates was structured using a fuzzy-coding technique, and a co-inertia analysis between traits and abundance data was used to interpret functional differences in bromeliad ecosystems. 4. The vegetative traits of A. mertensii depended on seed dispersion by C. femoratus and P. goeldii along a gradient of local conditions. The ant partner selected sets of invertebrates with traits that were best adapted to the bromeliads' morphology, and so the composition of the biological traits of invertebrate phytotelm communities depends on the identity of the ant partner. Biological traits suggest a bottom-up control of community structure in C. fenaoratus-associated phytotelmata and a greater structuring role for predatory invertebrates in P. goeldii-associated plants. 5. This study presents new information showing that two-species mutualisms affect the functional diversity of a much wider range of organisms. Most biological systems form complex networks where nodes (e.g. species) are more or less closely linked to each other, either directly or indirectly, through intermediate nodes. Our observations provide community-level information about biological interactions and functional diversity, and perspectives for further observations intended to examine whether large-scale changes in interacting species/community structure over broad geographical and anthropogenic gradients affect ecosystem functions.
Functional Ecology

hal-01096909
https://hal-agroparistech.archives-ouvertes.fr/hal-01096909
DOI : 10.1111/j.1365-2435-2011.01863.x

Éditeur(s) : HAL CCSD Elsevier
Résumé : International audience
ISSN: 0362-546X

hal-00699203
https://hal.archives-ouvertes.fr/hal-00699203

Éditeur(s) : HAL CCSD
Résumé : International audience
We propose an algorithm based on a model of visual perception that is meant to reflect the human judgment about the similarity of handwritten samples. The algorithm builds upon the Fuzzy Feature Contrast model and proposes an implementation of such a model in the domain of handwriting. The algorithm has been validated on the RIMES dataset, by comparing its performance with those of a panel of human experts. The experimental results show that the performance of the proposed algorithm is almost indistinguishable from the expert one and therefore may be a viable tool for handwriting comparison.
17th Biennial Conference of the International Graphonomics Society
Pointe-à-Pitre, Guadeloupe

hal-01165766
https://hal.univ-antilles.fr/hal-01165766
https://hal.univ-antilles.fr/hal-01165766/document
https://hal.univ-antilles.fr/hal-01165766/file/IGS_2015_submission_5.pdf

Résumé : La forte variabilité de la vitesse du vent fait que l'énergie produite par un parc éolien n'est pas constante dans le temps. Le gestionnaire ne peut donc pas dimensionner son réseau électrique en prenant intégralement ce type de production en compte. L' une des solutions préconisées pour permettre le développement de l' éolien et son intégration avec une plus grande sureté aux réseaux, est de développer et d'améliorer les outils de prévisions. Le travail de thèse consiste à améliorer les performances d'un outil de prédiction basé sur les réseaux de neurones bayesiens, permettant la prédiction de la puissance à très court terme . Le prédicteur fonctionne notamment par J'ajustement de paramètres, certain se détermine « automatiquement » via le mécanisme des réseaux de neurones bayesiens d' autres, que nous nommerons paramètres temporels, sont à l' appréciation de l'utilisateur. Le travail mené consiste à établir un protocole pour la fixation de ces paramètres tout en améliorant les performances du prédicteur . Nous avons donc décidé de conditionner leurs valeurs en fonction de la variabilité des séquences de puissance précédent l'instant de prévision. Tout d'abord nous avons classifié des séquences de puissance en fonction de leurs coefficients de variation en appliquant la méthode des C-moyennes floues. Puis, chaque classe formée a été testée sur plusieurs valeurs de paramètres, les valeurs associées aux meilleures prédictions ont été retenues. Enfin, ces résultats couplés au formalisme des Chaines de Markov, par le biais de la matrice de transition , ont perm is d'obtenir des taux d'amélioration par rapport à la persistance allant de 7,73 à 23,22 % selon l'horizon de prédiction considéré
The high variability of the wind speed has for conse quences that the energy produced by a wind farm is not constant over time. Therefore, the manager can't size the electrical network by takin g into account this type of production. One solution advocated for the development of wind energy and its integrati on with greater security at network, is to develop and improve fore casting tools. The thesi s objective is to improve the performance of a predi ction tool based on Bayesian neural networks, allowing the predi ction of wind power for short timescales. The predictor works, in part icular by the adjustment of parameters, sorne is determined "automatically" through the mechan ism of neural networks Bayesian other , which we cali temporal parameters are at the discretion of the user. The work involves establishing a protocol for the determination of these parameters and improving the performance of the predictor. So, we decided to condition their values depending on the sequence variability of wind power previous the moment of the forecast. First we classified sequences of power according to their coefficients of variation using the method of fuzzy C-means. Then, each formed class was tested for several parameters values, the values associated with the best predictions were selected. Finally , these result s coupled with the formalism of Markov chains , through the transition matrix allowed to obtain rates of improvement over the persistence ranging from 7.73 to 23.22 % depending on the prediction horizon considered

http://www.theses.fr/2013AGUY0679/document

Éditeur(s) : HAL CCSD Wiley
Résumé : International audience
Our understanding of the contribution of interspecific interactions to functional diversity in nature lags behind our knowledge of spatial and temporal patterns. Although two-species mutualisms are found in all types of ecosystems, the study of their ecological influences on other community members has mostly been limited to third species, while their influence on entire communities remains largely unexplored. 2. We hypothesized that mutualistic interactions between two respective ant species and an epiphyte mediate the biological traits composition of entire invertebrate communities that use the same host plant, thereby affecting food webs and functional diversity at the community level. 3. Aechmea mertensii (Bromeliaceae) is both a phytotelm ('plant-held water') and an ant-garden epiphyte. We sampled 111 bromeliads (111 aquatic invertebrate communities) associated with either the ant Pachycondyla goeldii or Camponotus femoratus. The relationships between ants, bromeliads and invertebrate abundance data were examined using a redundancy analysis. Biological traits information for invertebrates was structured using a fuzzy-coding technique, and a co-inertia analysis between traits and abundance data was used to interpret functional differences in bromeliad ecosystems. 4. The vegetative traits of A. mertensii depended on seed dispersion by C. femoratus and P. goeldii along a gradient of local conditions. The ant partner selected sets of invertebrates with traits that were best adapted to the bromeliads' morphology, and so the composition of the biological traits of invertebrate phytotelm communities depends on the identity of the ant partner. Biological traits suggest a bottom-up control of community structure in C. femoratus-associated phytotelmata and a greater structuring role for predatory invertebrates in P. goeldii-associated plants. 5. This study presents new information showing that two-species mutualisms affect the functional diversity of a much wider range of organisms. Most biological systems form complex networks where nodes (e. g. species) are more or less closely linked to each other, either directly or indirectly, through intermediate nodes. Our observations provide community-level information about biological interactions and functional diversity, and perspectives for further observations intended to examine whether large-scale changes in interacting species/community structure over broad geographical and anthropogenic gradients affect ecosystem functions.
ISSN: 0269-8463

hal-00817509
https://hal.archives-ouvertes.fr/hal-00817509