Éditeur(s) : HAL CCSD
Résumé : International audience
This article proposes a new indicator for handwriting examination in text independent framework. Experiments of writer identification are addressed using only handwritten short lines instead of a whole character. After preprocessing such as binarization and contours extraction, profiles of contours of handwritten lines were decomposed with fifth scales wavelet decomposition. As the result, we obtained indicators which showed qualities of handwritten lines such as smooth or jaggy. The indicators were analysed with Principle Component Analysis (PCA), and eigen vectors were obtained. In a phase of writer identification, using Kernel Orthogonal Mutual Subspace Method (KOMSM), subspace was calculated by the eigen vectors. The result obtained through the experiments was not enough to satisfy. In future works, the proposed method will be applied to whole handwritten characters.
17th Biennial Conference of the International Graphonomics Society
Pointe-à-Pitre, Guadeloupe

hal-01165761
https://hal.univ-antilles.fr/hal-01165761
https://hal.univ-antilles.fr/hal-01165761/document
https://hal.univ-antilles.fr/hal-01165761/file/IGS_2015_submission_2.pdf

Éditeur(s) : HAL CCSD Elsevier
Résumé : We present the hierarchical matrix (H-matrix) technique combined with the adaptive cross-approximation (ACA) applied to a three-dimensional (3D) elastostatic problem using the boundary element method (BEM). This is used in structural geology and geomechanics for the evaluation of the deformation and perturbed stress field associated with surfaces of displacement discontinuity. Such optimization significantly reduces (i) the time and memory needed for the resolution of the system of equations, but more importantly (ii) the time needed for the post-processing at observation points where the deformation and the perturbed stress field are evaluated. Specifically, it is shown that the H-matrix structure used with the ACA, clearly captures the kernel smoothness during the post-processing stage according to the field point positions, and optimizes the computation accordingly. Combined with the parallelization on multi-core processors, this technique allows intensive computations to be done on personal desktop and laptop computers. Numerical simulations are presented, showing the advantages of such optimizations compared to the standard method.
ISSN: 0955-7997

hal-00561032
https://hal.archives-ouvertes.fr/hal-00561032
DOI : 10.1016/j.enganabound.2009.10.016

Éditeur(s) : HAL CCSD Springer
Résumé : Third International Workshop, MLMI 2012, Held in Conjunction with MICCAI 2012
International audience
Universal Nearest Neighbours (unn) is a classifier recently proposed, which can also effectively estimates the posterior probability of each classification act. This algorithm, intrinsically binary, requires the use of a decomposition method to cope with multiclass problems, thus reducing their complexity in less complex binary subtasks. Then, a reconstruction rule provides the final classification. In this paper we show that the application of unn algorithm in conjunction with a reconstruction rule based on the posterior probabilities provides a classification scheme robust among different biomedical image datasets. To this aim, we compare unn performance with those achieved by Support Vector Machine with two different kernels and by a k Nearest Neighbours classifier, and applying two different reconstruction rules for each of the aforementioned classification paradigms. The results on one private and five public biomedical datasets show satisfactory performance.
Machine Learning in Medical Imaging
Nice, France

hal-00958909
https://hal.archives-ouvertes.fr/hal-00958909
https://hal.archives-ouvertes.fr/hal-00958909/document
https://hal.archives-ouvertes.fr/hal-00958909/file/Miccai.pdf
DOI : 10.1007/978-3-642-35428-1_15

Éditeur(s) : HAL CCSD
Résumé : International audience
In this paper, we use a multifractal approach based on the computation of two spectrums for image analysis and texture segmentation problems. The two spectrums are the Legendre Spectrum, determined by classical methods, and the Large Deviation Spectrum, determined by kernel density estimation. We propose a way for the fusion of these two spectrums to improve textured image segmentation results. An unsupervised k-means is used as clustering approach for the texture classification. The algorithm is applied on mosaic image built using IKONOS images and various natural textures from the Brodatz album. The segmentation obtained with our approach gives better results than the application of each spectrum separately.
CIARP Proceedings
Cancun, Mexico

hal-00602268
https://hal.archives-ouvertes.fr/hal-00602268