Olivier BernierOrange Labs
Olivier Bernier
About
59
Publications
15,427
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
1,061
Citations
Citations since 2017
Introduction
Publications
Publications (59)
Service Function Chaining (SFC) has gained momentum as one of the building blocks of the 5G ecosystem. Indeed, SFC combined with other promising technologies such as Software Defined Networking (SDN) as well as Network Slicing form the basis for enabling the 5G network services and fulfilling their requirements. Many contributions have been made to...
Service Function Chaining (SFC) is a networking concept by which traffic is steered through a set of ordered functions composing an end-to-end service. It represents one of the facilitating technologies for 5G, and is enabled by the Network Function Virtualization (NFV) and Software Defined Networks (SDN) paradigms. In the multi-domain context, SFC...
Counterfactual before counterfactual ? :-) - This chapter presents a new method to analyze the link between the probabilities produced by a classification model and the variation of its input values. The goal is to increase the predictive probability of a given class by exploring the possible values of the input variables taken independently. The p...
This article proposes a statistical approach for fast articulated 3D body tracking, similar to the loose-limbed model, but using the factor graph representation and a fast estimation algorithm. A fast Nonparametric Belief Propagation on factor graphs is used to estimate the current marginal for each limb. All belief propagation messages are represe...
The present work presents a novel framework for 3D content-based search and retrieval. On contrary to most state-of-the-art approaches, the query model can be not only an existing object from a database but also a model manually generated by the user using sketches. In the context of the proposed framework, three interfaces to the sketch-based 3D s...
Recently many methods for human articulated body tracking were proposed in the literature. These techniques are often computationally intensive and cannot be used for Human-Computer Interface. We propose in this article a real-time algorithm for upper body tracking with occultation handling. The tracking is based on an articulated body model, also...
Within the framework of collaborative interactions with 3D numerical copies of real objects inserted in virtual environments,
this paper tackles the issue of 3D object reconstruction from multiple calibrated cameras. After examining the various constraints
related to collaborative systems, we propose comprehensive, state of the art 3D reconstructio...
Recently many methods for human articulated body tracking were proposed in the literature. These techniques are often computationally
intensive and cannot be used for Human-Computer Interface. We propose in this article a real-time algorithm for upper body
tracking with occultation handling. The tracking is based on an articulated body model, also...
We present a method to analyse the properties of a synchronous stochastic neural network with asymmetric connections. We obtain an equivalence between the dynamics of the network and the iteration defined by a deterministic continuous mapping. This approach enables us to predict the existence of a bifurcation from a stable fixed point to an attract...
This paper describes a method for articulated 3D upper body tracking in monocular scenes using a graphical model to represent an articulated body structure. Belief propagation on factor graphs is used to compute the margi- nal probabilities of limbs. The body model is a loose-limbed model including attraction factors between adjacent limbs and cons...
This paper presents a new approach for automatic image color correction, based on statistical learning. The method both parameterizes color independently of illumination and corrects color for changes of illumination. This is useful in many image processing applications, such as image segmentation or background subtraction. The motivation for using...
This paper describes a method for articulated upper body tracking in monocular scenes. The compatibility between model and the image is estimated using one particle filter for each limb and the compatibility between limbs is represented by interaction potentials. The joint probability is obtained by belief propagation on a factor graph. The body mo...
SHIVA (Several-Humans Interface with Vision and Audio) est une interface multiutilisateurs, non intrusive, d'interaction libre par le geste et la parole avec de grands écrans. La tête et les mains de chaque personne sont suivies en temps réel à partir d'une caméra stéréoscopique. A partir de la position 3D de ces parties du corps, le système déterm...
This paper presents a new approach for automatic image color correction, based on statistical learning. The method both parameterizes
color independently of illumination and corrects color for changes of illumination. The motivation for using a learning approach
is to deal with changes of lighting typical of indoor environments such as home and off...
Virtual reality interfaces can immerse users into virtual environments from an impressive array of application fields, including entertainment, education, design, and navigation. However, history teaches us that no matter how rich the content is from these applications, it remains out of reach for users without a physical way to interact with it. M...
Among gestures naturally performed by users during communication, pointing gestures can be easily recognized and included
in more natural new Human Computer Interfaces. We approximate the eye-finger pointing direction of a user by detecting and
tracking, in real time, the 3D positions of the centre of the face and of both hands; the positions are o...
This paper presents a novel interactive framework for 3D content-based search and retrieval using as query model an object that is dynamically sketched by the user. In particular, two approaches are presented for generating the query model. The first approach uses 2D sketching and symbolic representation of the resulting gestures. The second utiliz...
This paper presents a new statistical approach for learning automatic color image correction. The goal is to parameterize color inde- pendently of illumination and to correct color for changes of illumination. This is useful in many image processing applications, such as color im- age segmentation or background subtraction. The motivation for using...
This article proposes a new statistical model for fast 3D articulated body tracking, similar to the loose-limbed model, but where inter-frame coherence is taken into ac- count by using the previous marginal probability of each limb as prior information. Belief propagation is used to estimate the current marginal for each limb. All probabil- ity dis...
This article proposes a new statistical model for fast 3D articulated body tracking, similar to the loose-limbed model, but using the factor graph repre- sentation. Belief propagation is used to estimate the current marginal for each limb. All belief propagation messages are represented as sums of weighted samples. The resulting algorithm correspon...
Constant background hypothesis for background subtraction algorithms is often not applicable in real environments because of shadows, reflections, or small moving objects in the background: flickering screens in indoor scenes, or waving vegetation in outdoor ones. In both indoor and outdoor scenes, the use of color cues for background segmentation...
In addition to being invariant to image rotation and translation, histograms have the advantage of being easy to compute. These advantages make histograms very popular in computer vision. However, without data quantization to reduce size, histograms are generally not suitable for realtime applications. Moreover, they are sensitive to quantization e...
Although large displays could allow several users to work together and to move freely in a room, their associated interfaces are limited to contact devices that must generally be shared. This paper describes a novel interface called SHIVA (Several-Humans Interface with Vision and Audio) allowing several users to interact remotely with a very large...
In this paper, we address the problem of the recognition of isolated complex mono- and bi-manual hand gestures. In the proposed system, hand gestures are represented by the 3D trajectories of blobs. The blobs are obtained by tracking colored body parts in real-time using the EM algorithm. In most of the studies on hand gestures, only small vocabula...
Among gestures naturally performed by users during communication, pointing gestures can be easily recognized and included in more natural new Human Computer Inter-faces. We approximate the eye-finger pointing direction of a user by detecting and tracking, in real time, the 3d positions of the centre of the face and of both hands; the positions are...
In this paper, we address the problem of the recognition of isolated complex mono- and bi-manual hand gestures. In the proposed system, hand gestures are represented by the 3D trajectories of blobs obtained by tracking colored body parts. In this paper, we study the results obtained on a complex database of mono- and bi-manual gestures. These resul...
Cet article présente une base de seize gestes dynamiques obtenus par suivi de différentes parties colorées du corps, suivi réalisé en temps réel par l'algorithme EM. Ces gestes réalisés avec une ou deux mains sont appris et reconnus avec des HMM. Les prétraitements effectués sur cette base de gestes ainsi que les résultats de classification sont pr...
In an image, the faces of the persons are the first information looked for. Performing efficient face detection in a video with persons (excluding cartoons and nature videos) allows to classify shots, and to obtain automatically face summaries. Shot sampling greatly improves time processing. Scene layout (same number of person, similar face positio...
Detecting faces in images with complex backgrounds is a difficult
task. Our approach, which obtains state of the art results, is based on
a neural network model: the constrained generative model (CGM).
Generative, since the goal of the learning process is to evaluate the
probability that the model has generated the input data, and constrained
since...
This paper presents a method for real time hand and head tracking,
in three dimensions, using two cameras. This tracking is intended as a
first step for a gesture recognition system, using the trajectories of
the hands, or as input to a real time clone animation system. The method
used is based on simple preprocessing followed by the use of a
stati...
The invention concerns an automatic system for sound and image recording in particular for videoconference, comprising means controlling (20) image and sound recording sensors (10) and sequence analysing means (40) monitoring said control means (20) to obtain automatic framing of the sequence being filmed. The invention is characterised in that an...
A new method to maximize the margin of MLP classier in classication problems is described. This method is based on a new,cost function which minimizes the variance of the mean,squared error. We show that with this cost function the generalization performance increase. This method is tested and compared,with the standard mean square error and is app...
A new hand gesture recognition method based on input-output hidden
Markov models is presented. This method deals with the dynamic aspects
of gestures. Gestures are extracted from a sequence of video images by
tracking the skin-color blobs corresponding to the hand into a body-face
space centered on the face of the user. Our goal is to recognize two...
Detecting faces in images with complex backgrounds is a difficult task. Our approach, which obtains state-of-the-art results, is based on a generative neural network model: the constrained generative model (CGM). To detect side-view faces and to decrease the number of false alarms, a conditional mixture of networks is used. To decrease the computat...
A novel method is presented to enhance the generalisation performances of multi-layer perceptron used as discriminant networks. We clearly show how to modify the learning criterium in order to control the error distribution. This control allows to obtain a better margin in classification problems. Results are given for two different problems. These...
Realtime face detection
We propose a model for image space discretisation based on face location and on body anthropometry. In this body-face space,
a neural network recognizes hand postures. The neural network is a constrained generative model already applied to face detection.
We present a generating-function-based method to analyse the dynamics of synchronous neural networks. It is applied to a very general model of networks of automata, which includes a certain number of neural network models of the Little type as particular cases. The method enables us to analyse the evolution of the order parameters as well as their...
In a face to face meeting, participants are free to choose who
they look at and who they listen to. In a videoconference, participants
can only see or hear the audio-visual signal that the distant
participants have decided or have been able to transmit. We present
Panorama, a contactless visual man machine interface which allows to
explore a distan...
A real-time system is described for automatic detection and
tracking of multiple persons, in the context of video-conferencing
systems. This system, called MULTRAK (multiperson locating and tracking
automatic kernel) is able to continuously detect and track the position
of faces in its field of view. The heart of the system as a modular
neural netw...
We present a neural network approach to human face detection. Using a modular system, a conditional mixture of networks, we are able to detect front view faces as well as turned faces (up to 50 degrees) with excellent performances. This modular network is integrated into LISTEN, our face tracking system. It enables this system to detect and track i...
![FIG. 1 — [a] : Image d'origine issue d'une séquence. [b] : Détection de...](profile/Michel-Collobert/publication/236151819/figure/fig1/AS:340777914912768@1458259194025/a-Image-dorigine-issue-dune-sequence-b-Detection-de-la-teinte-de-la-peau_Q320.jpg)
Les informations visuelles et acoustiques sont au coeur de la (télé)communication entre les personnes. Le visage est la principale source d'information. Des techniques de détection du mouvement et de la teinte de la peau délimitent des régions d'intérêt où peuvent se trouver des visages. Un réseau de neurones détecte le visage et fournit la positio...
A generative neural network model, constrained by non-face examples chosen by an iterative algorithm, is applied to face detection. To improve the generalization ability of the model, another constraint based on the minimum description length is added. This model is tested and compared with another state-of-the-art face detection system on a large...
A new learning model based on autoassociative neural networks is developped and applied to face detection. To extend the detection ability in orientation and to decrease the number of false alarms, different combinations of networks are tested: ensemble, conditional ensemble and conditional mixture of networks. The use of a conditional mixture of n...
Both visual and acoustical informations provide effective means of telecommunication between persons. In this context, the face is the most important part of the person both visually and acoustically. We describe how the cooperation of image and audio processing allows to track a person's face and to collect the audio information it produces. We pr...
Résumé Cet article présente une méthode de détection de la tête d'une personne dans une scène, par transformée de Hough elliptique. Pour cela, nous proposons une nouvelle transformée de Hough plus rapide. Plutôt que d'accumuler dans l'espace de Hough entier, l'accumulation se fait dans deux zones rectangulaires de l'espace de Hough, en fonction de...
This paper presents a new method to analyze the link between the probabilities produced by a classification model and the variation of its input values. The goal is to increase the predictive probability of a given class by exploring the possible values of the input variables taken independently. The proposed method is presented in a general framew...
Résumé Cet article décrit une méthode destinée au suivi du haut du corps en 3D pour des scènes filmées avec une caméra monoculaire. La structure articulaire est représentée par un modèle graphique probabiliste, à savoir un graphe de facteurs dans lequel la propagation des croyances permet d'évaluer la probabilité marginale de chacun des membres. Le...
The present report presents the framework and the results of Project 7: "A Multimodal (Gesture+Speech) Interface for 3D Model Search and Retrieval Integrated in a Virtual Assembly Application", which has been developed during the eNTERFACE-2005 summer workshop in the context of the SIMILAR NoE. The "MASTER-PIECE" (Multimodal Assembly with SIMILAR T...
Computer vision based man machine interfaces allows the user to freely interact. In order to do so, it is necessary to detect and track user's body parts and to analyze their movements. Hands are our most common mean of inter- action with the world. The face is the body part the most characteristic of a human being. This paper deals with the simult...
This article proposes a new method for D reconstruction of real world objects using a low number of views. The method uses a hierarchical octree representation of a 3D voxel space. An iterative algorithm is used : starting from a coarse resolution, precise 3D models are obtained. The contributions of this paper concern a new algorithm for estimatin...
France Telecom -CNET DTL / DLI / TNT 2 avenue Pierre Marzin 22307 Lannion FRANCE Les interfaces gestuelles basées sur l'image sont la voie la plus naturelle pour la construction des interfaces homme-machine du futur, mais ce sont aussi les plus difficiles à concevoir. Nous proposons d'utiliser un modèle de réseau de neurones pour reconnaître une po...
La méthode d'analyse du flot optique dans les séquences d'images animées que nous proposons est basée sur l'utilisation d'un réseau d'automates. Cette méthode permet d'introduire naturellement l'existence de frontières dans le flot optique, dues à la présence de plusieurs objets animés de mouvements différents dans la séquence.
Projects
Projects (2)
coming soon...
A PhD will start on this topic (October 2017).
The general context is the security and the cybercrime (detection of the frauds). Given these concerns, system to detect frauds, abnormal behaviors or signatures (of attacks) are among the major components of the current devices to protect systems. The attempts of frauds, intrusions or attacks are protean, evolutionary and made furtive by the simple Big Data volumetric of the traffic and/or the transactions.
The supervision, of a network or transactions requires the consideration, on one hand the specificity of deceitful acts and on the other hand constraints connected to the data which are varied, voluminous and with high velocity.
The approach of detection by comparison to a database of behavior said normal, for example by means of rules, presents limits. The establishment of rules is a complex work, because they have to follow constantly the evolutions of the known frauds. Besides, the unlisted deceitful detection of acts is not possible in essence with this method.
Face to these points, IA methods, using learning machines, widen the potential of fraud detection, known or unknowns, and can be associated with technologies or mechanisms relative to Big Data.