Serge Miguet | Université Lumière Lyon 2 (original) (raw)

Papers by Serge Miguet

Research paper thumbnail of A multi-resolution approach for a content-based image retrieval on the grid

HAL (Le Centre pour la Communication Scientifique Directe), Jan 29, 2004

Hospitals and medical centers are producing more and more data, that need to be pro­ cessed. Thos... more Hospitals and medical centers are producing more and more data, that need to be pro­ cessed. Those data are confidential, heterogeneous, and limited to the geographic site where they have been produced. Unless properly anonymised, they cannot be distributed on wide area networks. Grid technologies allow the globalization of storage and processing resources, and enable large-scale experimentations on distributed data. They constitute a promising tool to treat the different data and analyse the knowledge they contain, while offering secured access and high-performance computing capacities to the different users. Our aim is to evaluate the possibilities of grid technologies for handling medical data. In this paper, we focus on a breast cancer diagnosis assistance tool, based on a distributed and incremental knowledge construction and a content-based image retrieval system. We analyse the different scenarii of uses of such a tool. We further propose an algorithm that indexes mammographie images for content-based query purposes. This algorithm is tested on images of different resolutions in or der to reduce the indexation time and we analyse its performances with first experiments on the grid.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Indexation d'images mammographiques par les ondelettes de Gabor

HAL (Le Centre pour la Communication Scientifique Directe), Oct 29, 2008

Bookmarks Related papers MentionsView impact

Research paper thumbnail of <title>Optimized and load-balanced portable parallel Zbuffer</title>

Proceedings of SPIE, Apr 7, 1995

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Three-Dimensional Imaging of the Wrist

This paper presents a technique offering a three-dimensional segmented representation of the bone... more This paper presents a technique offering a three-dimensional segmented representation of the bones of the wrist from a set of CT-slices. Many authors have studied the way of computing images of 3D data by so called “volume rendering” techniques. The major drawback of these approaches is that they cannot process independently the different objects appearing in the volume. Our method separates the objects, allowing to render surfaces hidden by other objects or to evaluate precisely the location and the shape of a fracture. Since we want to be able to process each bone independently, we apply a segmentation step to the set of slices, producing a set of contours. In order to classify these contours into different bones, we use a proximity criterion based on the computation of their best approximation ellipses. This gives a wireframe modelization of the bones. The external surfaces are then approximated by triangulation or spline interpolation and visualized by classical rendering algorithms.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Détection de Deepfakes par Réseaux de Convolution : Performances et Limites Actuelles

HAL (Le Centre pour la Communication Scientifique Directe), Sep 6, 2022

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Scattering on on a ring of processors

HAL (Le Centre pour la Communication Scientifique Directe), 1990

International audienceno abstrac

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Invariant Model Combining Geometry and Appearance for Facial Detection and Gender Classification From Arbitrary Viewpoints

Advances in multimedia and interactive technologies book series, Jul 13, 2017

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Complexity of discrete surfaces in the Dividing-cubes algorithm

Springer eBooks, 1996

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Real Time Foreground-Background Segmentation Using a Modified Codebook Model

Real time segmentation of scene into objects and background is really important and represents an... more Real time segmentation of scene into objects and background is really important and represents an initial step of object tracking. Starting from the codebook method we propose some modifications which show significant improvements in most of the normal and also difficult conditions. We include parameter of frequency for accessing, deleting, matching and adding codewords in codebook or to move cache codewords into codebook. We also propose an evaluation method in order to objectively compare several segmentation techniques, based on receiver operating characteristic (ROC) analysis and on precision and recall method. We propose to summarize the quality factor of a method by a single value based on a weighted Euclidean distance or on a harmonic mean between two related characteristics.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Discrete Geometry for Computer Imagery: 6th International Workshop, DGCI'96, Lyon, France, November 13 - 15, 1996, Proceedings

On recent trends in discrete geometry in computer science.- Least squares fitting of digital poly... more On recent trends in discrete geometry in computer science.- Least squares fitting of digital polynomial segments.- Geometrical parameters extraction from discrete paths.- A 3D-hole closing algorithm.- Multiresolution representation of shape in binary images.- Discrete elastica.- Fractal representation of planar shapes.- Volume synthesis.- Fast computation of the normal vector field of the surface of a 3-D discrete object.- Finite element meshes by means of voxels.- Polyhedra generation from lattice points.- Recognizing arithmetic straight lines and planes.- A topologically consistent representation for image analysis: The Frontiers Topological Graph.- Determining the components of the complement of a Digital (n?1)-manifold in ?n.- Determination of discrete sampling grids with optimal topological and spectral properties.- Inter-pixel Euclidean paths for image analysis.- Filling driven by contour marching.- Hypergraph model of digital topology for grey level images.- Application of skeletonization algorithms for myocardial spect quantification.- Supercover 3D Polygon.- Topologically defined isosurfaces.- Iso-surface extraction in 4D with applications related to scale space.- Complexity of discrete surfaces in the Dividing-cubes algorithm.- Circle digitization and cellular automata.- Reconstructing convex polyominoes from horizontal and vertical projections II.- The Quasi-Shear rotation.- Computational geometry and discrete computations.- Some topological properties of discrete surfaces.- A statistical process for surface tracking.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Application of Large Scale Particle Image Velocimetry to measure surface velocity field in a stormewater detention basin and to validate CFD model

HAL (Le Centre pour la Communication Scientifique Directe), Sep 10, 2017

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Proceedings of the 6th International Workshop on Discrete Geometry for Computer Imagery

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Reduction operations on a distributed memory machine with a reconfigurable interconnection network

Bookmarks Related papers MentionsView impact

Research paper thumbnail of A new data structure for efficient search on isovists

arXiv (Cornell University), Nov 10, 2022

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Mobile Augmented Reality : Optimization methods for pose estimation and geometric measurements

HAL (Le Centre pour la Communication Scientifique Directe), Dec 1, 2022

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Automatic Traffic analysis in video sequences

Third International Conference on Advances In Computing, Control And Networking - ACCN 2015, 2015

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Multiple Object Tracking on Smart Embedded Camera For Automated Conveying Systems

Proceedings of the 11th International Conference on Distributed Smart Cameras, 2017

In this paper we present an embedded smart cameras system that performs multiple objects tracking... more In this paper we present an embedded smart cameras system that performs multiple objects tracking for the control of an automated conveying system. The smart cameras automatically extract scene description models of the conveying environment from underlying motion structure, as observed over time. In particular, the ROI mask and the conveyor's pathways model are used to improve the tracking performance during on-line inference. An efficient parallelization strategy of background subtraction is proposed. Our approach takes advantage of multi-core embedded platforms. We provide an evaluation of the proposed method on videos coming from an industrial application.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Analyse de l'activité humaine dans les séquences vidéo

Cet expose se propose de presenter quelques unes des contributions recentes de l’equipe Imagine d... more Cet expose se propose de presenter quelques unes des contributions recentes de l’equipe Imagine du LIRIS, dans le domaine de l’analyse de l’activite humaine par la video. Plusieurs de ces contributions ont ete developpees conjointement avec la jeune entreprise Foxstream, incubee par le LIRIS au sein de l’Universite Lumiere Lyon 2, et specialisee dans l’edition de logiciels d’analyse video intelligente. Dans le domaine de la segmentation d’objets en mouvements, nous nous sommes interesses a l’optimisation de la methode des Codebooks, et proposons une methodologie d’evaluation des algorithmes de segmentation premier plan / arriere plan. Nous presentons une technique de suivi d’objets en mouvement dans les videos, specialisee dans le suivi des humains, et basee sur l’elaboration d’un modele d’apparence articule. Pour rendre le modele plus robuste aux occultations partielles des membres, nous proposons une approche basee sur un appariement d’un modele 3D avec la silhouette 2D obtenue da...

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Three dimensional discrete morphing: Application to medical atlas construction

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Invariant Model Combining Geometry and Appearance for Facial Detection and Gender Classification From Arbitrary Viewpoints

In this chapter, we tackle in the same process the problems of face detection and gender classifi... more In this chapter, we tackle in the same process the problems of face detection and gender classification, where the faces present a wide range of the intra-class appearance are taken from arbitrary viewpoints. We try to develop complete probabilistic model to represent and learn appearance of facial objects in both shape and geometry with respect to a landmark in the image, and then to be able to predict presence and position of the appearance of the studied object class in new scene. After have predicted the facial appearance and the geometry of invariants, geometric hierarchical clustering combines different prediction of positions of face invariant. Then, the algorithm of cluster selection with a best appearance localizes faces in the image. Using a probabilistic classification, each facial feature retained in the detection step will be weighted by a probability to be male or female. This set of features contributes to determine the gender associated to a detected face. This model...

Bookmarks Related papers MentionsView impact

Research paper thumbnail of A multi-resolution approach for a content-based image retrieval on the grid

HAL (Le Centre pour la Communication Scientifique Directe), Jan 29, 2004

Hospitals and medical centers are producing more and more data, that need to be pro­ cessed. Thos... more Hospitals and medical centers are producing more and more data, that need to be pro­ cessed. Those data are confidential, heterogeneous, and limited to the geographic site where they have been produced. Unless properly anonymised, they cannot be distributed on wide area networks. Grid technologies allow the globalization of storage and processing resources, and enable large-scale experimentations on distributed data. They constitute a promising tool to treat the different data and analyse the knowledge they contain, while offering secured access and high-performance computing capacities to the different users. Our aim is to evaluate the possibilities of grid technologies for handling medical data. In this paper, we focus on a breast cancer diagnosis assistance tool, based on a distributed and incremental knowledge construction and a content-based image retrieval system. We analyse the different scenarii of uses of such a tool. We further propose an algorithm that indexes mammographie images for content-based query purposes. This algorithm is tested on images of different resolutions in or der to reduce the indexation time and we analyse its performances with first experiments on the grid.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Indexation d'images mammographiques par les ondelettes de Gabor

HAL (Le Centre pour la Communication Scientifique Directe), Oct 29, 2008

Bookmarks Related papers MentionsView impact

Research paper thumbnail of <title>Optimized and load-balanced portable parallel Zbuffer</title>

Proceedings of SPIE, Apr 7, 1995

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Three-Dimensional Imaging of the Wrist

This paper presents a technique offering a three-dimensional segmented representation of the bone... more This paper presents a technique offering a three-dimensional segmented representation of the bones of the wrist from a set of CT-slices. Many authors have studied the way of computing images of 3D data by so called “volume rendering” techniques. The major drawback of these approaches is that they cannot process independently the different objects appearing in the volume. Our method separates the objects, allowing to render surfaces hidden by other objects or to evaluate precisely the location and the shape of a fracture. Since we want to be able to process each bone independently, we apply a segmentation step to the set of slices, producing a set of contours. In order to classify these contours into different bones, we use a proximity criterion based on the computation of their best approximation ellipses. This gives a wireframe modelization of the bones. The external surfaces are then approximated by triangulation or spline interpolation and visualized by classical rendering algorithms.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Détection de Deepfakes par Réseaux de Convolution : Performances et Limites Actuelles

HAL (Le Centre pour la Communication Scientifique Directe), Sep 6, 2022

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Scattering on on a ring of processors

HAL (Le Centre pour la Communication Scientifique Directe), 1990

International audienceno abstrac

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Invariant Model Combining Geometry and Appearance for Facial Detection and Gender Classification From Arbitrary Viewpoints

Advances in multimedia and interactive technologies book series, Jul 13, 2017

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Complexity of discrete surfaces in the Dividing-cubes algorithm

Springer eBooks, 1996

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Real Time Foreground-Background Segmentation Using a Modified Codebook Model

Real time segmentation of scene into objects and background is really important and represents an... more Real time segmentation of scene into objects and background is really important and represents an initial step of object tracking. Starting from the codebook method we propose some modifications which show significant improvements in most of the normal and also difficult conditions. We include parameter of frequency for accessing, deleting, matching and adding codewords in codebook or to move cache codewords into codebook. We also propose an evaluation method in order to objectively compare several segmentation techniques, based on receiver operating characteristic (ROC) analysis and on precision and recall method. We propose to summarize the quality factor of a method by a single value based on a weighted Euclidean distance or on a harmonic mean between two related characteristics.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Discrete Geometry for Computer Imagery: 6th International Workshop, DGCI'96, Lyon, France, November 13 - 15, 1996, Proceedings

On recent trends in discrete geometry in computer science.- Least squares fitting of digital poly... more On recent trends in discrete geometry in computer science.- Least squares fitting of digital polynomial segments.- Geometrical parameters extraction from discrete paths.- A 3D-hole closing algorithm.- Multiresolution representation of shape in binary images.- Discrete elastica.- Fractal representation of planar shapes.- Volume synthesis.- Fast computation of the normal vector field of the surface of a 3-D discrete object.- Finite element meshes by means of voxels.- Polyhedra generation from lattice points.- Recognizing arithmetic straight lines and planes.- A topologically consistent representation for image analysis: The Frontiers Topological Graph.- Determining the components of the complement of a Digital (n?1)-manifold in ?n.- Determination of discrete sampling grids with optimal topological and spectral properties.- Inter-pixel Euclidean paths for image analysis.- Filling driven by contour marching.- Hypergraph model of digital topology for grey level images.- Application of skeletonization algorithms for myocardial spect quantification.- Supercover 3D Polygon.- Topologically defined isosurfaces.- Iso-surface extraction in 4D with applications related to scale space.- Complexity of discrete surfaces in the Dividing-cubes algorithm.- Circle digitization and cellular automata.- Reconstructing convex polyominoes from horizontal and vertical projections II.- The Quasi-Shear rotation.- Computational geometry and discrete computations.- Some topological properties of discrete surfaces.- A statistical process for surface tracking.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Application of Large Scale Particle Image Velocimetry to measure surface velocity field in a stormewater detention basin and to validate CFD model

HAL (Le Centre pour la Communication Scientifique Directe), Sep 10, 2017

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Proceedings of the 6th International Workshop on Discrete Geometry for Computer Imagery

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Reduction operations on a distributed memory machine with a reconfigurable interconnection network

Bookmarks Related papers MentionsView impact

Research paper thumbnail of A new data structure for efficient search on isovists

arXiv (Cornell University), Nov 10, 2022

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Mobile Augmented Reality : Optimization methods for pose estimation and geometric measurements

HAL (Le Centre pour la Communication Scientifique Directe), Dec 1, 2022

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Automatic Traffic analysis in video sequences

Third International Conference on Advances In Computing, Control And Networking - ACCN 2015, 2015

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Multiple Object Tracking on Smart Embedded Camera For Automated Conveying Systems

Proceedings of the 11th International Conference on Distributed Smart Cameras, 2017

In this paper we present an embedded smart cameras system that performs multiple objects tracking... more In this paper we present an embedded smart cameras system that performs multiple objects tracking for the control of an automated conveying system. The smart cameras automatically extract scene description models of the conveying environment from underlying motion structure, as observed over time. In particular, the ROI mask and the conveyor's pathways model are used to improve the tracking performance during on-line inference. An efficient parallelization strategy of background subtraction is proposed. Our approach takes advantage of multi-core embedded platforms. We provide an evaluation of the proposed method on videos coming from an industrial application.

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Analyse de l'activité humaine dans les séquences vidéo

Cet expose se propose de presenter quelques unes des contributions recentes de l’equipe Imagine d... more Cet expose se propose de presenter quelques unes des contributions recentes de l’equipe Imagine du LIRIS, dans le domaine de l’analyse de l’activite humaine par la video. Plusieurs de ces contributions ont ete developpees conjointement avec la jeune entreprise Foxstream, incubee par le LIRIS au sein de l’Universite Lumiere Lyon 2, et specialisee dans l’edition de logiciels d’analyse video intelligente. Dans le domaine de la segmentation d’objets en mouvements, nous nous sommes interesses a l’optimisation de la methode des Codebooks, et proposons une methodologie d’evaluation des algorithmes de segmentation premier plan / arriere plan. Nous presentons une technique de suivi d’objets en mouvement dans les videos, specialisee dans le suivi des humains, et basee sur l’elaboration d’un modele d’apparence articule. Pour rendre le modele plus robuste aux occultations partielles des membres, nous proposons une approche basee sur un appariement d’un modele 3D avec la silhouette 2D obtenue da...

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Three dimensional discrete morphing: Application to medical atlas construction

Bookmarks Related papers MentionsView impact

Research paper thumbnail of Invariant Model Combining Geometry and Appearance for Facial Detection and Gender Classification From Arbitrary Viewpoints

In this chapter, we tackle in the same process the problems of face detection and gender classifi... more In this chapter, we tackle in the same process the problems of face detection and gender classification, where the faces present a wide range of the intra-class appearance are taken from arbitrary viewpoints. We try to develop complete probabilistic model to represent and learn appearance of facial objects in both shape and geometry with respect to a landmark in the image, and then to be able to predict presence and position of the appearance of the studied object class in new scene. After have predicted the facial appearance and the geometry of invariants, geometric hierarchical clustering combines different prediction of positions of face invariant. Then, the algorithm of cluster selection with a best appearance localizes faces in the image. Using a probabilistic classification, each facial feature retained in the detection step will be weighted by a probability to be male or female. This set of features contributes to determine the gender associated to a detected face. This model...

Bookmarks Related papers MentionsView impact