Gilles Venturini

• Pr, Dr, Ing,
• Professor at University of Tours

In this paper, we study the visual mining of time series, and we contribute to the study and evaluation of 3D tubular visualizations. We describe the state of the art in the visual mining of time-dependent data, and we concentrate on visualizations that use a tubular shape to represent data. After analyzing the motivations for studying such a repre...

In this paper, we study how to visualize large amounts of multidimensional data with a radial visualization. For such a visualization, we study a multi-threaded implementation on the CPU and the GPU. We start by reviewing the approaches that have visualized the largest multidimensional datasets and we focus on the approaches that have used CPU or G...

Introduction: Diffusion tractography relies on complex mathematical models that provide anatomical information indirectly, and it needs to be validated. In humans, up to now, tractography has mainly been validated by qualitative comparison with data obtained from dissection. No quantitative comparison was possible because Magnetic Resonance Imagin...

In this paper, we propose a new method for the visual reorganization of online analytical processing (OLAP) cubes that aims at improving their visualization. Our method addresses dimensions with hierarchically organized members. It uses a genetic algorithm that reorganizes k-ary trees. Genetic operators perform permutations of subtrees to optimize...

We propose a process to compress a pre-trained Vision Language Model into a ternary version of itself instead of training a ternary model from scratch. A new initialization scheme from pre-trained weights based on the k-means algorithm is proposed to reduce the ternarization time. We implement different custom operators for executing the ternary mo...

Current large open vision models could be useful for one and few-shot object recognition. Nevertheless, gradient-based re-training solutions are costly. On the other hand, open-vocabulary object detection models bring closer visual and textual concepts in the same latent space, allowing zero-shot detection via prompting at small computational cost....

Current artificial neural networks are trained with parameters encoded as floating point numbers that occupy lots of memory space at inference time. Due to the increase in the size of deep learning models, it is becoming very difficult to consider training and using artificial neural networks on edge devices. Binary neural networks promise to reduc...

We present in this paper an augmented reality system called VISIT designed to deliver content about artworks to visitors. This system aims to simplify the use of augmented reality by allowing (non-computer) users to define content and place it in 3D around an artwork. VISIT is composed of two main generic elements with respect to historical or arti...

We present in this paper the state of the art and an analysis of recent research work and achievements performed in the domain of AI-based and vision-based systems for helping blind and visually impaired people (BVIP). We start by highlighting the recent and tremendous importance that AI has acquired following the use of convolutional neural networ...

We study in this work the properties of a new method called Gen-POIViz for data projection and visualization. It extends a radial visualization with a genetic-based optimization procedure so as to find the best possible projections. It uses as a basis a visualization called POIViz that uses Points of Interest (POIs) to display a large dataset. This...

Convolutional neural networks (CNNs), in a few decades, have outperformed the existing state of the art methods in classification context. However, in the way they were formalised, CNNs are bound to operate on euclidean spaces. Indeed, convolution is a signal operation that are defined on euclidean spaces. This has restricted deep learning main use...

Convolutional neural networks (CNNs), in a few decades, have outperformed the existing state of the art methods in classification context. However, in the way they were formalised, CNNs are bound to operate on euclidean spaces. Indeed, convolution is a signal operation that are defined on euclidean spaces. This has restricted deep learning main use...

Insects are living beings whose utility is critical in life sciences. They enable biologists obtaining knowledge on natural landscapes (for example on their health). Nevertheless, insect identification is time-consuming and requires experienced workforce. To ease this task, we propose to turn it into an image-based pattern recognition problem by re...

Many tasks in computer vision and pattern recognition are formulated as graph matching problems. Despite the NP-hard nature of such problems, fast and accurate approximations have led to significant progress in a wide range of applications. However, learning graph matching from observed data, remains a challenging issue. In practice, the node corre...

Thanks to the capturing devices cost reduction and the advent of social networks, the size of image collections is becoming extremely huge. Many works in the literature have addressed the indexing of large image collections for search purposes. However, there is a lack of support for exploratory data mining. One may want to wander around the images...

Entomology has had many applications in many biological domains (i.e insect counting as a biodiversity index). To meet a growing biological demand and to compensate a decreasing workforce amount, automated entomology has been around for decades. This challenge has been tackled by computer scientists as well as by biologists themselves. This survey...

Many tasks in computer vision and pattern recognition are formulated as graph matching problems. Despite the NP-hard nature of the problem, fast and accurate approximations have led to significant progress in a wide range of applications. Learning graph matching functions from observed data, however, still remains a challenging issue. This paper pr...

This book presents a collection of representative and novel work in the field of data mining, knowledge discovery, clustering and classification, based on expanded and reworked versions of a selection of the best papers originally presented in French at the EGC 2014 and EGC 2015 conferences held in Rennes (France) in January 2014 and Luxembourg in...

Ever-growing image collections are common in several �elds such as health, digital humanities or social networks. Nowadays, there is a lack of visualisation tools to browse such large image collection. In this work, the incremental indexing and the visualisation of large image collections is done jointly. The BIRCH algorithm is improved to incre- m...

Open data (OD) contributes to the spread and publication of life sciences data on the Web. Searching and filtering OD datasets, however, can be challenging since the metadata that accompany the datasets are often incomplete or even non-existent. Even when metadata are present and complete, interpretation can be complicated owing to the quantity, va...

Open data (OD) contributes to the spread and publication of life sciences data on the Web. Searching and filtering OD datasets, however, can be challenging since the metadata that accompany the datasets are often incomplete or even non-existent. Even when metadata are present and complete, interpretation can be complicated owing to the quantity, va...

Dans cet article, nous étudions de manière conjointe la construction et l’exploration visuelle d’une structure de classification pour de très grande base d’images. Pour garantir que la structure construite vérifiera les contraintes de taille nécessaires à sa visualisation dans une interface Web tout en reflétant les propriétés topologiques des donn...

Nous présentons dans cet article une nouvelle méthode de classification non-supervisée appelée AntClust, inspirée du système de reconnaissance chimique des fourmis. Celui-ci, connu sous le nom de fermeture coloniale, repose sur l'apprentissage et le partage d'une odeur coloniale communè a toutes les fourmis d'un même nid. Dans notre méthode, une fo...

Data integration has always been a major problem in computer sciences. The more heterogeneous, large and distributed the data sources become, the more difficult the data integration process is. Nowadays, more and more information is being made available on the Web. This is especially the case in the Open Data (OD) movement. Large quantities of data...

This paper addresses the problem of the incremental con-struction of an indexing structure, namely a proximity graph, for largeimage collections. To this purpose, a local update strategy is examined.Considering an existing graph G and a new node q, how only a relevantsub-graph of G can be updated following the insertion of q? For a givenproximity g...

Nowadays, more and more information is flowing in and is provided on the Web. Large datasets are made available covering many fields and sectors. Open Data (OD) plays an important role in this field. Thanks to the volumes and the variety of the released datasets, OD brings high societal and business potential. In order to realize this potential, th...

In the Fourth edition of "Big Data Mining and Visualization" Days

We study in this work how a user can be guided to find a relevant visualization in the context of visual data mining. We present a state of the art on the user assistance in visual and interactive methods. We propose a user assistant called VizAssist, which aims at improving the existing approaches along three directions: it uses simpler computatio...

New and different information sources have appeared over the past years (e.g. Blogs, Media, Open Data, Scientific Data and Social Networks). The variety of these sources is growing and the related data volume does not cease to increase exponentially. Open Data (OD) initiatives and platforms are one of the current major data producers, also because...

A la 15ème Conférence Internationale sur l'Extraction et la Gestion des Connaissances (EGC) - Atelier VIF

For several years, and even decades, data integration has been a major problem in computer sciences. When it becomes necessary to process information from different data sources, several problems may appear, making the process of integration more difficult. Nowadays, more and more information is being sent and received and is made available on the...

In this paper, we propose an original solution to the problem of point cloud clustering. The proposed technique is based on a d-dimensional formulated Delaunay Triangulation (DT) construction algorithm and adapts it to the problem of cluster detection. The introduced algorithm allows this detection as along with the DT construction. Precisely, a cr...

In this paper we study a 3D tubular visualization of software activity log data, with the aim of supporting multithreaded software development and debugging. We consider an existing visualization called Datatube2 that has already been used to help various domain experts in the analysis of large amounts of time-dependent data. Since software logs ar...

http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=6901365

This book is a collection of representative and novel works done in Data Mining, Knowledge Discovery, Clustering and Classification that were originally presented in French at the EGC'2012 Conference held in Bordeaux, France, on January 2012. This conference was the 12th edition of this event, which takes place each year and which is now successful...

This paper presents an original feature vector extraction process based on the Delaunay triangulation (DT) and a zoning technique. The presented work provides an illustration of the equivalency between a zoning and the Delaunay triangulation in the context of handwritten character recognition. A novel technique that relies on the approximation of a...

We deal in this paper with the problem of creating an interactive and visual map for a large collection of Open datasets. We first describe how to define a representation space for such data. We use text mining techniques to create features. Then, with a similarity measure between Open datasets, we use the Relative Neighbors method for building a p...

We deal in this paper with the problem of creating an interactive and visual map for a large collection of Open datasets. We first describe how to define a representation space for such data, using text mining techniques to create features. Then, with a similarity measure between Open datasets, we use the k-nearest neighbors method for building a p...

We study in this paper the use of a 3D interface for OLAP. We analyze the state of the art in OLAP 3D visualizations. In a first step, we propose a new interface, called VR4OLAP, that combines the relative advantages of the studied methods. VR4OLAP can visualize 3 dimensions of an OLAP data cube and up to two measures. It represents the OLAP operat...

We deal in this paper with the problem of creating an interactive and visual map for a large collection of Open datasets. We first describe how to define a representation space for such data, using text mining techniques to create features. Then, with a similarity measure between Open datasets, we use the K-nearest neighbors method for building a p...

This paper presents an immersive visualization tool that helps anatomists to establish a ground truth for brain white matter fiber bundles. Each step of a progressive anatomical dissection of human brain hemisphere is acquired using a high resolution 3D laser scanner and a photographic device. Each resulting surface is textured with a high resoluti...

The recent and novel research contributions collected in this book are extended and reworked versions of a selection of the best papers that were originally presented in French at the EGC’2011 Conference held in Brest, France, on January 2011. EGC stands for "Extraction et Gestion des connaissances" in French, and means "Knowledge Discovery and Man...

We present in this paper a tool called EXOD (“EXploration of Open Datasets”) for the visual analysis of a large collection of open datasets. EXOD aims at helping the users to find datasets of interest. EXOD starts with the download of a large collection of datasets from an Open data web site. For each dataset, it extracts its meta-data and its cont...

In this paper, we propose to parallelize, on CPU and GPU, a radial-based visualization method that uses points of interests.We show that this approach may visualize in a few seconds millions of data with tens of dimensions, and we study the efficiency of the parallel approach in different configurations.

In this paper we present a summary of our work which has led to the conception of a new model for the fast construction of proximity graphs. We present the state of the art in graph self-assembly, and then we detail the self-assembly behavior observed in real ants from which our model is derived. We describe our main algorithm, called AntGraph, whe...

We deal in this paper with the problem of automating the process of choosing an appropriate visualization and its parameters in the context of visual data mining (VDM). To solve this problem, we develop a user assistant that performs 2 steps: the system starts by suggesting to users different mappings between their data and possible visualizations....

In this paper, a new visual and interactive user interface for OLAP is presented, and its strengths and weaknesses examined. A survey on 3D interfaces for OLAP is detailed, which shows that only one interface that uses Virtual Reality has been proposed. Then we present our approach: it consists of a 3D representation of OLAP cubes where many OLAP o...

We deal in this paper with the problem of automating the process of choosing a visualization and its parameters in visual data mining. To solve this problem, we have developed a user assistant that performs 2 steps: the system starts by suggesting to users different matchings between their database and the possible visualizations. These matchings a...

We deal in this paper with the problem of automating the process of choosing a visualization and its parameters in visual data mining. To solve this problem, we have developed a user assistant that performs 2 steps: the system starts by suggesting to users different matchings between their database and the possible visualizations. These matchings a...

We present in this paper a new incremental and bio-inspired algorithm that builds proximity graphs for large amounts of data (i.e. 1 million). It is inspired from the self-assembly behavior of real ants where each ant progressively becomes attached to an existing support and then successively to other attached ants. The ants that we have defined wi...

We deal in this paper with the problem of automating the process of choosing a visualization and its parameters in data mining. To solve this problem, we have developed a user assistant that performs 2 steps: the system start by suggesting to users different matchings between their database and a set of the visualizations. These matchings are gener...

We present a system to keep track of a destructive process such as a medical specimen dissection, from data acquisition to interactive and immersive visualization, in order to build ground truth models. Acquisition is a two-step process, first involving a 3D laser scanner to get a 3D surface, and then a high resolution camera for capturing the text...

INTRODUCTION Despite variations depending on acquisition and post-treatment parameters, MRI tractography is imperfectly validated. 4 causes may explain this incomplete validation, and especially the lack of quantitative approach: (1) for ethical reasons, and even if imperfect, only postmortem techniques can be used in human as “silverstandard”. Amo...

Nous nous intéressons dans cet article à la fouille visuelle de données temporelles, où les données ont été mises sous la forme de n attributs dont les valeurs sont enregistrées pendant k instants. Après un état de l'art sur les différentes approches de visualisation de telles séries, nous présentons plus particulièrement une approche ayant reçue e...

Outils pour l'acquisition et le calcul d'images stéréoscopiques, le calibrage de caméras (couleurs et calibrage stéréoscopique), le calcul de profondeurs, la visualisation d'images en stéréoscopie (mesures interactives 3D et annotation d'images 3D), la gestion d'un ensemble d'images 3D et le partage de connaissances entre experts

We present in this paper a new method for the visual and interactive exploration of Web sites logs. Web usage data is mapped onto a 3D tube which axis represents time and where each facet corresponds to the hits of a given page and for a given time interval. A rearrangement clustering algorithm is used to create groups among pages. Several interact...

We present in the paper a system that integrates all hardware and software to extract information from 3D images of skin. It is composed of a lighting equipment and stereoscopic cameras, a camera calibration algorithm that uses evolutionary principles, virtual reality equipment to visualize the images and interact with them in 3D, a set of interact...

We present in this paper a new method for building and exploring a 3D hypermedia in virtual reality and for a biomedical domain. Starting from the acquisition of stereoscopic images and from the calibration of cameras, our system offers the user the possibility to visualize these images in 3D and to annotate specific areas with texts or voice recor...

We present in the paper a system called Skin3D that integrates all hardware and software to extract information from 3D images of skin. It is composed of a lighting equipment and acquisition-based stereoscopic cameras, a camera calibration using genetic algorithms, virtual reality equipment to restore the images and interact in 3D with them, a set...

Plateforme logicielle pour la fouille de données visuelle et interactive en réalité virtuelle