Éditeur(s) : Edinburgh, London
Siècle(s) traité(s) : 19
Provenance : Université Caen Normandie. Service commun de la documentation
Permalien : http://www.manioc.org/patrimon/CAE19033_3
CAE19033_3

Éditeur(s) : London : Saunders and Otley
Description : History of the Caribs. Culture and way of life of Antiguans in the 18th and 19th Centuries.
Siècle(s) traité(s) : 18
Droits : Domaine public
Provenance : Ville de Pointe-à-Pitre. Réseau des bibliothèques
Permalien : http://www.manioc.org/patrimon/PAP11081
PAP11081

Éditeur(s) : Baltimore : Johns Hopkins Press
Siècle(s) traité(s) : 19
Droits : Domaine public
Provenance : Université Toulouse 1 Capitole. Service Commun de la Documentation
Permalien : http://www.manioc.org/patrimon/TLS17024
TLS17024

Résumé : Publiée dans le cadre du premier Congrès international d'histoire coloniale à Paris
Siècle(s) traité(s) : 20
Droits : Domaine public
Provenance : Ville de Pointe-à-Pitre. Réseau des bibliothèques
Permalien : http://www.manioc.org/patrimon/PAP11103
PAP11103

Éditeur(s) : Londres : John Stockolale
Siècle(s) traité(s) : 18
Droits : Domaine public
Provenance : Archives départementales de Guadeloupe
Permalien : http://www.manioc.org/patrimon/ADG17063
ADG17063

Éditeur(s) : Edinburgh, London
Siècle(s) traité(s) : 19
Provenance : Université Caen Normandie. Service commun de la documentation
Permalien : http://www.manioc.org/patrimon/CAE19033_4
CAE19033_4

Éditeur(s) : HAL CCSD
Résumé : International audience
This paper introduces the information criteria in feature selection framework. Information criteria are integrated in feature selection scheme to select subset candidates. The accuracy of the proposed approach is based on the quality of the probability density approximations of these features. They are obtained using histograms optimized thanks to the adaptive arithmetic coding principles. Tests on simulated data and references are made. Multiple classifiers are used. The correct classification rate shows, the importance of this tools and its ability to select a best subsets. Generally, the subsets produce a good characterization of classes which the data belong.
TAIMA Proceedings

hal-00634754
https://hal.archives-ouvertes.fr/hal-00634754

Éditeur(s) : Boston : Estes and Lauriat
Siècle(s) traité(s) : 19
Provenance : Archives départementales de Guadeloupe
Permalien : http://www.manioc.org/patrimon/ADG18205
ADG18205

Éditeur(s) : London : Hamilton, Adams & Co
Siècle(s) traité(s) : 19
Droits : Domaine public
Provenance : Ville de Pointe-à-Pitre. Réseau des bibliothèques
Permalien : http://www.manioc.org/patrimon/HASH0103f0f83da4829c83c1b88d
HASH0103f0f83da4829c83c1b88d

Éditeur(s) : HAL CCSD
Résumé : International audience
The k-nearest neighbors (k-NN) classification rule is still an essential tool for computer vision applications, such as scene recognition. However, k-NN still features some major drawbacks, which mainly reside in the uniform voting among the nearest prototypes in the feature space. In this paper, we propose a new method that is able to learn the "relevance" of prototypes, thus classifying test data using a weighted k-NN rule. In particular, our algorithm, called Multi-class Leveraged k-nearest neighbor (MLNN), learns the prototype weights in a boosting framework, by minimizing a surrogate exponential risk over training data. We propose two main contributions for improving computational speed and accuracy. On the one hand, we implement learning in an inherently multiclass way, thus providing significant computation time reduction over one-versus-all approaches. Furthermore, the leveraging weights enable effective data selection, thus reducing the cost of k-NN search at classification time. On the other hand, we propose a kernel generalization of our approach to take into account real-valued similarities between data in the feature space, thus enabling more accurate estimation of the local class density. We tested MLNN on three datasets of natural images. Results show that MLNN significantly outperforms classic k-NN and weighted k-NN voting. Furthermore, using an adaptive Gaussian kernel provides significant performance improvement. Finally, the best results are obtained when using MLNN with an appropriate learned metric distance.
Proceedings of the 10th Asian Conference on Computer Vision, ACCV 2010, November 8-12, 2010, Queenstown, New Zealand
Queenstown, New Zealand

hal-00664606
https://hal.inria.fr/hal-00664606
https://hal.inria.fr/hal-00664606/document
https://hal.inria.fr/hal-00664606/file/mlnn_accvfinal.pdf

Éditeur(s) : HAL CCSD
Résumé : International audience
The large majority of methods proposed in literature for handwriting recognition assume that words are produced drawing large parts of the ink without lifting the pen, other than horizontal bars and dots. This fundamental assumption, however, does not always hold: while some educational systems provide explicit training for producing continuous handwriting, minimizing the number of pen-up during the production of a word, others do not. As a consequence, whenever the handwriting presents pen-up within a word, the recognition performance can drop significantly. In a preliminary study, we presented an algorithm for discriminating among different types of ink appearing in handwriting, namely isolated characters, cursive, dots, horizontal and vertical bars, based on the use of a suitable set of features. In this paper, we have characterized the discriminative power of each considered feature according to different measures and we have proposed a method for combining the different feature rankings. We have also used the Fischer's Linear Discriminant Analysis (LDA) for exhaustively selecting the best feature subsets with increasing number of features. Finally, we have compared the results obtained by using the feature subsets provided by LDA with those obtained with the feature subsets selected according to our feature ranking. The experimental results, on different datasets of handwritten words, showed that our approach successfully achieves its aim allowing to reduce the computational cost without affecting the overall performance of the recognition process.
17th Biennial Conference of the International Graphonomics Society
Pointe-à-Pitre, Guadeloupe

hal-01165877
https://hal.univ-antilles.fr/hal-01165877
https://hal.univ-antilles.fr/hal-01165877/document
https://hal.univ-antilles.fr/hal-01165877/file/IGS_2015_submission_13.pdf

Éditeur(s) : New-York : American anti-slavery society
Siècle(s) traité(s) : 18
Droits : Domaine public
Provenance : Collectivité territoriale de Martinique. Bibliothèque Schoelcher
Permalien : http://www.manioc.org/patrimon/SCH13214
SCH13214

Éditeur(s) : London : William Kidd
Siècle(s) traité(s) : 19
Droits : Domaine public
Provenance : Archives départementales de Guadeloupe
Permalien : http://www.manioc.org/patrimon/ADG17046
ADG17046

Éditeur(s) : HAL CCSD Springer Verlag
Résumé : 17 pages, 11 figures.
International audience
The problem addressed in this paper is the automatic extraction of names from a document image. Our approach relies on the combination of two complementary analyses. First, the image-based analysis exploits visual clues to select the regions of interest in the document. Second, the textual-based analysis searches for name patterns and low-level word textual features. Both analyses are then combined at the word level through a neural network fusion scheme. Reported results on degraded documents such as facsimile and photocopied technical journals demonstrate the interest of the combined approach.
ISSN: 1433-7541

hal-00329459
https://hal.archives-ouvertes.fr/hal-00329459
DOI : 10.1007/s10044-006-0038-6

Éditeur(s) : HAL CCSD
Résumé : National audience
Designing systems able to interact with students in a natural manner is a complex and far from solved problem. A key aspect of natural interaction is the ability to understand and appropriately respond to human emotions. This paper details our response to the continuous Audio/Visual Emotion Challenge (AVEC'12) whose goal is to predict four affective signals describing human emotions. The proposed method uses Fourier spectra to extract multi-scale dynamic descriptions of signals characterizing face appearance, head movements and voice. We perform a kernel regression with very few representative samples selected via a supervised weighted-distance-based clustering, that leads to a high generalization power. We also propose a particularly fast regressor-level fusion framework to merge systems based on different modalities. Experiments have proven the efficiency of each key point of the proposed method and our results on challenge data were the highest among 10 international research teams.
IHM'14, 26e conférence francophone sur l'Interaction Homme-Machine
Lille, France

hal-01089628
https://hal.archives-ouvertes.fr/hal-01089628
https://hal.archives-ouvertes.fr/hal-01089628/document
https://hal.archives-ouvertes.fr/hal-01089628/file/p44-nicole.pdf

Éditeur(s) : HAL CCSD AGU and the Geochemical Society
Résumé : International audience
The Ganos fault is the westernmost segment of the North Anatolian Fault that experienced the Mw = 7.4 earthquake of 9 August 1912. The earthquake revealed 45-km-long of surface ruptures inland, trending N70 E, and 5.5 m of maximum right lateral offset near Güzelköy. The long-term deformation of the fault is clearly expressed by several pull-apart basins and sag ponds, pressure and shutter ridges and offset streams. In parallel with detailed geomorphologic investigations, we measured co-seismic and cumulative displacements along the fault, and selected the Güzelköy site for paleoseismology. A microtopographic survey at the site yields 10.5 0.5 m and 35.4 1.5 m cumulative lateral offsets of stream channels and geomorphologic features. Seven paleoseismic parallel and cross-fault trenches document successive faulting events and provide the timing of past earthquakes on the Ganos fault segment. Radiocarbon dating of successive colluvial wedges in trench T1, and the fresh scarplet above (probably 1912 surface rupture) indicate the occurrence of three faulting events since the 14th century. Parallel trenches (3, 5, 6 and 7) expose paleo-channels and show a cumulative right-lateral offset of 16.5 1.5 m next to the fault, and 21.3 1.5 m total channel deflection. Radiocarbon dating of past channel units and fault scarp-related colluvial deposits imply an average 17 +/ 5 mm/year slip rate and 323 142 years recurrence interval of large earthquakes during the last 1000 years on the Ganos fault. The succession of past faulting events and inferred slip rate west of the Marmara Sea provide more constraint on the long-term faulting behavior in the seismic gap of the North Anatolian Fault and may contribute to a better seismic hazard assessment in the Istanbul region.
ISSN: 1525-2027

hal-00748218
https://hal.archives-ouvertes.fr/hal-00748218
DOI : 10.1029/2011GC003960

Éditeur(s) : Londres : John Stockolale
Siècle(s) traité(s) : 18
Droits : Domaine public
Provenance : Archives départementales de Guadeloupe
Permalien : http://www.manioc.org/patrimon/ADG17064
ADG17064

Éditeur(s) : London : Richard Phillips
Siècle(s) traité(s) : 19
Droits : Domaine public
Provenance : Archives territoriales de Guyane
Permalien : http://www.manioc.org/patrimon/GAD12003
GAD12003

Éditeur(s) : HAL CCSD BioMed Central
Résumé : International audience
UNLABELLED: ABSTRACT: BACKGROUND: Malaria remains a major health problem in French Guiana, with a mean of 3800 cases each year. A previous study in Camopi, an Amerindian village on the Oyapock River, highlighted the major contribution of environmental features to the incidence of malaria attacks. We propose a method for the objective selection of the best multivariate peridomestic landscape characterisation that maximises the chances of identifying relationships between environmental features and malaria incidence, statistically significant and meaningful from an epidemiological point of view. METHODS: A land-cover map, the hydrological network and the geolocalised inhabited houses were used to characterise the peridomestic landscape in eleven discoid buffers with radii of 50, 100, 200, 300, 400, 500, 600, 700, 800, 900 and 1000 metres. Buffer-based landscape characterisations were first compared in terms of their capacity to discriminate between sites within the geographic space and of their effective multidimensionality in variable space. The Akaike information criterion (AIC) was then used to select the landscape model best explaining the incidences of P. vivax and P. falciparum malaria. Finally, we calculated Pearson correlation coefficients for the relationships between environmental variables and malaria incidence, by species, for the more relevant buffers. RESULTS: The optimal buffers for environmental characterisation had radii of 100 m around houses for P. vivax and 400 m around houses for P. falciparum. The incidence of P. falciparum malaria seemed to be more strongly linked to environmental features than that of P. vivax malaria, within these buffers. The incidence of P. falciparum malaria in children was strongly correlated with proportions of bare soil (r = -0.69), land under high vegetation (r = 0.68) and primary forest (r = 0.54), landscape division (r = 0.48) and the number of inhabited houses (r = -0.60). The incidence of P. vivax malaria was associated only with landscape division (r = 0.49). CONCLUSIONS: The proposed methodology provides a simple and general framework for objective characterisation of the landscape to account for field observations. The use of this method enabled us to identify different optimal observation horizons around houses, depending on the Plasmodium species considered, and to demonstrate significant correlations between environmental features and the incidence of malaria.
ISSN: 1476-072X

inserm-00674075
http://www.hal.inserm.fr/inserm-00674075
http://www.hal.inserm.fr/inserm-00674075/document
http://www.hal.inserm.fr/inserm-00674075/file/1476-072X-10-65.pdf
DOI : 10.1186/1476-072X-10-65

Éditeur(s) : HAL CCSD
Résumé : International audience
MICCAI Workshop
Boston, United States

hal-01144677
https://hal.archives-ouvertes.fr/hal-01144677