Topics of interest (most of them being hereabove illustrated)

Image understanding using a priori structural and qualitative knowledge (since 2004)
Implementation of image processing algorithms :

Software (since 2002)
Hardware (from 1998 to ~2006)

Applications :

Medical image analysis (since 1998)

Computer aided surgery (from 2002 to 2009)
Estimation of blood flow in carotid from Phase Contrast MRI (from 2010 to 2015)

IoT for healthcare (since 2015)
Document analysis (from 1998 to 2002)

Image understanding : spatial relationships, graph and hypergraph matching

I have been involved in the proposal of a method exploiting spatial relationships and hypergraph matching for postnatal piglet brain extraction, in the context of the study of the maturation process of the infant brain. Due to the small size of the piglet brain and the abundance of surrounding fat and muscles, the automatic brain extraction using correctely initialized deformable models is tedious, and the standard approach used for human brain does not apply. The paper proposes an original brain extraction method based on a deformable model, whose initialization is guided by a priori known spatial relationships between some anatomical structures of the head. This concerns a structural model related to a priori known relative position of eyes, nose and brain center, assumed to be an anatomical invariant. Using this structural model, our proposal detects both eyes and nose, from which one deduces the brain center, for finally initializing deformable models. Anatomical structures are retrieved by matching observed relationships with those embedded in the a priori structural model. This hypergraph matching, where hypergraph matching concerns relative position of eyes and nose (ternary constraint related to these 3 entities). The method has been implemented and preliminary experiments have been performed on a set of 6 piglets, to evaluate the accuracy of the brain center localization, the one of the final brain extraction using deformable models. The brain center is correctly localized with a mean error of 1.7 mm, underlying the relevance of the approach.

Principle of the proposed approach

Hypergraph matching

A. Durandeau, J.-B. Fasquel, I. Bloch, E Mazerand, P Menei, C. Montero-Menei, M. Dinomais, Structural information and hypergraph matching for MRI piglet brain extraction, 10th International Conference on Pattern Recognition Systems, Tours, France, 2019 [link]

Image understanding : inclusion and photometric relationships, inexact graph matching

I have been involved in developping a method for recovering and identifying image regions from an initial oversegmentation using qualitative knowledge of its content. This method focuses on simple a priori qualitative inclusion and photometric relationships such as ”region B is included in region A”, ”the intensity of region A is lower than the one of region B” or ”regions A and B depict similar intensities” (photometric uncertainty). The proposed method is based on a two steps’ inexact graph matching approach (search for the best subgraph isomorphism and then merge of remaining regions). This method have been evaluated on various images (grayscale and color, medical or not), with various intial oversegmentation algorithms (k-means, meanshift, quickshift). Hereabove are reported some figures illustrating this approach. Both code and data are freely available.

Principle of the proposed approach

Example : application to a medical image

Example : application to a color image (with photometric uncertainty)

J.-B. Fasquel, N. Delanoue,A graph based image interpretation method using a priori qualitative inclusion and photometric relationships, IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, IF: 9.455 (2017) [link][Code-Database] [github]

Code : Scikit-gtimage github, zip

Documentation

Database

Image understanding : inclusion & photomety, sequential analysis, k-means, exact graph matching

I have been interesting in exploiting, within a sequential interpretation procedure, a priori known qualitative inclusion and photometric relationships between image regions. A sequential interpretation procedure assumes that regions are progressively segmented and recognized by associating them with corresponding nodes in graphs related to the prior qualitative knowledge. The main contribution concerns the parameterization of the k-means clustering algorithm, to be used during the segmentation procedure, and the exact-graph-matching-based identification of resulting clusters, corresponding to regions declared in graphs. This approach has been evaluated real gray-scale and color images with dedicated sequential analysis procedures. Processing results have been compared with those obtained without the proposed parameterization of k-means, as well as with some other clustering approaches. Results showed the relevance of this approach, in particular in terms of segmentation accuracy, computation time, and seeding reliability.

Principle of the proposed approach

Example : application to a medical image

Example : application to a color image

J.-B. Fasquel, N. Delanoue, An approach for sequential image interpretation using a priori binary perceptual topological and photometric knowledge and k-means based segmentation, Journal of the Optical Society of America A, 2018, IF: 1.6 (2016), [link].

Image understanding : intersection relationships, abdominal structures, interactive sequential interpretation

This work regarded an interactive interpretation method for patient modeling in computer aided diagnosis. The proposal aimed at guiding the interactive sequential segmentation and identification of abdominal structures, according to a priori known intersection relationships between them. In particular, this approach allowed to dynamically define regions of interest (ROIs) within which computations can be constrained, regardless the segmentation strategy. The goal was then to avoid the processing of irrelevant image points, therefore improving and accelerating computations. In the case of a sequential modeling procedure, the proposed ROI management method enabled, for delineating a given medical structure, to optimally determine image points of interest by taking previously segmented structures into account. This method based on reasoning on the graph embedding both relationships and the knowledge of already identified medical structures.

J.B. Fasquel,V. Agnus, J. Moreau, L. Soler, J. Marescaux. An Interactive Medical Segmentation System Based on the Optimal Management of Regions of Interest Using Topological Medical Knowledge, Computer Methods and Programs in Biomedicine 2006 ; 216-230, Volume 82, Issue 3 [link,preprint], IF: 1.51

J.B. Fasquel, V. Agnus, L. Soler, J. Marescaux, A Hierarchical Topological Knowledge Based Image Segmentation Approach Optimizing the Use of Contextual Regions of Interest: Illustration for Medical Image Analysis, In Proceedings of the IEEE International Conference on Image Processing, Atlanta (USA), 08-11 Octobre 2006, p777-780

Application : Measuring the evolution of both lumen area and blood flow in carotid from Phase Contrast MRI

Phase-Contrast (PC) velocimetry magnetic resonance imaging (MRI) is a useful modality to explore cardiovascular pathologies, but requires the automatic segmentation of vessels and the measurement of both lumen area and blood flow evolutions. In this work, I have involved the proposal of a semi-automated method for extracting lumen boundaries of the carotid artery and compute both lumen area and blood flow evolutions over the cardiac cycle. This method used narrow band region-based active contours in order to correctly capture the lumen boundary without being corrupted by surrounding structures. This approach has been compared to traditional edgebased active contours, considered in related works, which significantly underestimate lumen area and blood flow. Experiments have been performed using both a sequence of a homemade phantom and sequences of 20 real carotids, including a comparison with manual segmentation performed by a radiologist expert. Results obtained on the phantom sequence showed that the edge-based approach leads to an underestimate of carotid lumen area and related flows of respectively 18:68 % and 4:95 %. This appeared significantly larger than weak errors obtained using the region-based approach (respectively 2.73 % and 1.23 %). Benefits appeared even better on the real sequences. The edge-based approach led to underestimates of 40.88 % for areas and 13.39 % for blood flows, compared to limited errors of 7.41 % and 4.6 % with our method. Experiments also illustrateed the high variability and therefore the lack of reliability of manual segmentation.

J.-B. Fasquel, A. Lécluse, C. Cavaro-Ménard, S. Willoteaux, A semi-automated method for measuring the evolution of both lumen area and blood flow in carotid from Phase Contrast MRI, Computers In Biology And Medicine, Vol. 66, 2015 [link] [preprint]

G. Trébuchet, J.-B. Fasquel, A. Lecluse, C. Cavaro Ménard, S. Willoteaux, Region-based active contours for computer-aided analysis of carotid Phase Contrast MRI, The 26th IEEE International Symposium on Computer-Based Medical Systems, Porto, Portugal, 20-22 June, 2013

G. Trébuchet, J.-B. Fasquel, C. Cavaro-Ménard, S. Willoteaux, Coupling anatomical and functional information for the computer-aided delineation of Phase-Contrast MRI images using active contours, IPTA 2012 : International Conference on Image Processing Theory, Tools and Applications, Istanbul, Turkey, 15-18 October, 2012

Application : patient modeling for computer aided surgery

Medical image processing leads to a major improvement of patient care by guiding the surgical gesture. From these initial data, new technologies of virtual reality and augmented reality can increase the potential of such images. The 3D modeling of patients from their CT-scan or MRI thus allows an improved surgical planning. Simulation offers the opportunity to train the surgical gesture before carrying it out. These two preoperative steps can be used intra-operatively thanks to the development of augmented reality (AR) which consists in superimposing the pre-operative 3D modeling of the patient onto the real intra-operative view of the patient. AR provides surgeons with a view in transparency of their patient and can also guide surgeons thanks to the virtual improvement of their real surgical tools that are tracked in real time during the procedure. During the intervention, augmented reality therefore offers surgeons a view in transparency of their patient, what tomorrow will lead to the automation of the most complex manoeuvres. In the near future, thanks to the exploitation of these systems, surgeons will program and check on the virtual clone of the patient an optimal procedure without errors, which will be replayed on the real patient by the robot under surgeon control. This medical dream used to be virtual, but today it is about to become reality. In this context, I have been mainly involved in conceiving algorithms (e.g. based on qualitative knwoledge) and tools (e.g. softwares, interactive algorithms) for modeling anatomical and pathological structures of patients (mainly of the abdominal area) from 3D medical images.

L. Soler, S. Nicolau, A. Hostettler, J.-B. Fasquel, V. Agnus, A. Charnoz, J. Moreau, B. Dallemagne, D. Mutter and J. Marescaux. Computer Assisted Digestive Surgery, M. Garbey et al. (Editors), Computational Surgery and Dual Training, ISBN 978-1-4419-1122-3, Springer, p. 139-153, 2010 [link,preprint]

J.B. Fasquel, G. Brocker, J. Moreau, V. Agnus, N. Papier, C. Koehl, L. Soler, J. Marescaux, A Modular and Evolutive Component Oriented Software Architecture for Patient Modeling, Computer Methods and Programs in Biomedecine, 2006, Volume 83, Issue 3 [link,preprint], IF: 1.51

Soler L, Nicolau S, Hostettler A, Fasquel JB, Agnus V, Charnoz A, Moreau J, Dallemagne B, Mutter D, and Marescaux J, Virtual Reality and Augmented Reality applied to endoscopic and NOTES procedures. IFMBE (Medical Physics and Biomedical Engineering) proceedings (springer ed.), 25(6):362-365, 2009. [link,preprint]

L. Soler, S. Nicolau, J.B. Fasquel, V. Agnus, A. Charnoz, A. Hostettler, J. Moreau, C. Forest, D. Mutter, J. Marescaux. Virtual reality and augmented reality applied to laparoscopic and notes procedures.Biomedical Imaging: From nano to macro, 2008, ISBI 2008, 5^th IEEE International Symposium on. 1399-1402 [link,preprint]

Software : example of IoT plateform in healthcare

I have been involved in the conception and the implementation of an in-home aggregation plateform for monitoring physiological parameters, and involving two objective physi- cal sensors (bio-impedanceter and thermometer) and a subjective one (fatigue level perceived by the patient). This plateform used modern IoT-related technologies such as embedded systems (Raspberry Pi and Arduino) and the publish/subscribe communication pattern (MQTT communication protocol in our case). Compared to many related works, monitoring was enterely achieved using a box as a central element, while the mobile device (tablet) was only used for controlling the acquisition procedure using a simple web browser, without any specific application.

A. Jamin, J.-B. Fasquel, M. Lhommeau, E. Cornet, S. Abadie-Lacourtoisie, S. Henni and G. Leftheriotis, An aggregation plateform for IoT-based healthcare: illustration for bioimpedancemetry, temperature and fatigue level monitoring, Internet of Things Technologies for HealthCare, Third International Conference, HealthyIoT 2016, Västerås, Sweden, October 18-19, 2016

Software : generic design and medical software product line

One of the challenges in software development regards the appropriate coupling of separated code elements in order to correctly build initially expected high-level software functionalities. In this context, I have been interested in addressing issues related to the dynamic composition of such code elements (i.e. how they are dynamically plugged together) as well as their collaboration (i.e. how they work together). I also considered the limitation of build-level dependencies, to avoid the entire re-compilation and re-deployment of a software when modifying it or integrating new functionalities. To solve these issues, a new design pattern has been proposed, coupling role and component concepts. The relevance of this pattern has been illustrated for medical software (image processing for patient modeling, visualization, augmented reality, interaction with surgical robot). Compared to most related works focusing on few role concepts while ignoring others, the proposed pattern integrated many role concepts as first-class entities, including in particular a refinement of the notion of collaboration. Another significant contribution of this proposal concerned the coupling of role and component concepts. Roles were related to the functional aspects of a target software program (composition and collaboration of functional units). Components corresponded to the physical distribution of code elements with limited build-level dependencies. As illustrated by this work, such a coupling enabled to instantiate a software program using a generic main program together with a description file focusing on software functionalities only. Related code elements were transparently retrieved and composed at run-time before appropriately collaborating, regardless the specificity of their distribution over components. Smiley face

J.-B. Fasquel,Johan Moreau, A design pattern coupling role and component concepts: Application to medical software, The Journal of Systems and Software, 84, 847-863, 2011 [link, preprint], IF: 1.293 (2010 JCR)

Fasquel J.B., Chabre G., Zanne P., Nicolau S., Agnus V., Soler L., De Mathelin, M., Marescaux J., A role-based component architecture for computer assisted interventions: illustration for electromagnetic tracking and robotized motion rejection in flexible endoscopy, Workshop on Systems & Architecture for Computer Assisted Intervention, the 12th International Conference on Medical Image Computing and Computer Assisted Intervention, London, MICCAI 2009 [link,preprint]

J.B. Fasquel, J. Waechter, S. Nicolau, V. Agnus, L. Soler, J. Marescaux, A XML based component oriented architecture for image guided surgery : illustration for the video based tracking of a surgical tool, Workshop on Systems Architecture for Computer Assisted Intervention, the 11th International Conference on Medical Image Computing and Computer Assisted Intervention, New York, MICCAI 2008 [link,preprint]

Code : github

Software : extended iterator design pattern for processing image subregions

I have been interested in proposing a design pattern to extend the ITK library capabilities (Insight ToolKit, supported by the National Library of Medicine) for processing arbitrary shaped regions of interest. The proposed design, which was based on an extension of the well-known iterator design pattern, allowed the processing of regions of interest with arbitrary shapes, without modifying the existing ITK code. This proposal has been experimentally evaluated by considering the practical case of the liver vessel segmentation from CT-scan images, where it appeared relevant to constrain processings to the liver area. Experimental results clearly proved the interest of this work: for instance, the anisotropic filtering of this area has been performed in only 16 s with this design, while it took 52 s using the native ITK framework. A major advantage of this method was that only add-ons were required, therefore facilitating the deployement of the pattern over the entire ITK libary, while preserving native core entities. Smiley face

J.B. Fasquel, V. Agnus and J. Lamy, An efficient and generic extension to ITK to process arbitrary shaped regions of interest, Computer Methods and Programs in Biomedecine, 2006, Volume 81, Issue 1, January [link,preprint], IF: 1.51

J.B. Fasquel and V. Agnus, Improving genericity and performances of medical systems and softwares based on image analysis, The 18^th IEEE International Symposium on Computer-Based Medical Systems, June 23-24, 2005, Dublin, 247 - 252

Hardware : Opto-electronic reconstruction of 3D medical images

After some PhD works on optical image processing, I have been involved, from 2003 to 2006, in a project focusing on the fast opto-electronic image reconstruction from tomodensitometric data provided by volume exploration systems (typically medical images from CT-Scanners). The potential of optical architectures has been compared with electronic/digital solutions, considering computation power and processed image quality. An opto-electronic system has then been proposed and evaluated (simulation and experiments), in particular for the computation of complex reconstruction algorithms, e.g. ASSR for helical CT-scanner.

M. Madec, W. Uhring, E. Hueber,J.-B. Fasquel, J. Bartinger, Y. Hervé, Methods for improvement of spatial light modulator image rendering, Optical Engineering, 48(3), 1, 2009 [link], IF: 0.722 (2008 JCR)

M. Madec., J.B. Fasquel, W. Uhring, P. Joffre, Y. Hervé, Optoelectronic implementation of helical cone-beam computed tomography algorithms, Optical Engineering 47, 2008 [link], IF: 0.757 (2007 JCR)

M. Madec., J.B. Fasquel, W. Uhring, P. Joffre, Y. Hervé, Optical implementation of the filtered backprojection algorithm, Optical Engineering 46, 2007 [link], IF: 0.885

M. Madec., W. Uhring, J.B. Fasquel, P. Joffre, Y. Hervé, Compatibility of temporal multiplexed spatial light modulator with optical image processing, Optics Communications. 275, 27-37, 2007 [link], IF: 1.35

J.B. Fasquel, M. Bruynooghe, A hybrid opto-electronic method for real-time automatic verification of handwritten signatures, International Journal for Document Analysis and Recognition, No. 1, March 2004, pp. 56-68 [link,preprint], IF: 0.4

J.B. Fasquel, M. Bruynooghe, New hybrid opto-electronic method for fast and unsupervised object detection, Optical Engineering, Vol. 42, Issue 11, p. 3352-3364, November 2003 [link,preprint], IF: 0.885

Supervision of students

PhD Students

2018-2021 : Franck Mouney, co-advised with T. Tiplica (30%) and M. Dinomais (supervisor)

Title: "Bracelet connecté pour la détection précoce d'accidents cardiovasculaires"

2010-2013 : Guillaume Trébuchet, co-advised with C. Cavaro-Ménard and S. Willoteaux (supervisor).

Title: "Segmentation par contours actifs de séquences de vélocimétrie IRM: Application aux artères carotides"

2003-2006 : Morgan Madec, co-advised with W. Uhring and Y. Hervé (supervisor).

Title: "Conception, simulation et réalisation d’un processeur optoélectronique pour la reconstruction d’images médicales" [pdf]
Current status: assistant professor at University of Strasbourg

Master Students

2018 : Colin Gachet, Master in Applied Mathematics ("Mathématiques et Applications, spécialité Systèmes Dynamiques et Signaux"), University of Angers.

Title : "Quelques contributions de la vision au domaine agricole"

2018 : Martin Schmidt, Master in Applied Mathematics ("Mathématiques et Applications, spécialité Systèmes Dynamiques et Signaux"), University of Angers.

Title : "Analyse de séries temporelles pour la détection de profil en gériatrie à partir d'un cycle connecté"

2017 : Antoine Jamin, Master in Applied Mathematics ("Mathématiques et Applications, spécialité Systèmes Dynamiques et Signaux"), University of Angers.

Title : "Suivi à domicile de patients traités pour un cancer à l'aide d'un tensiomètre connecté"

2011 : Christophe Rigaud, Master in Applied Mathematics ("Mathématiques et Applications, spécialité Systèmes Dynamiques et Signaux"), University of Angers.

Title : "Interprétation d'images et graphe conceptuel intégrant des connaissances topologiques et photométriques",pdf]
Current status: PhD student at La Rochelle University (L3i) [link]

2009 : Guillaume Chabre, Master in Computer Science ("Informatique, parcours Image et Connaissance"), University of Reims

Title: "Intégration logicielle générique d'un système chirurgical robotisé" [pdf]

2009 : Vivianne Leguy, Master in Mathematics ("Mathématiques, Spécialité Processus, Automatique, Simulation, Statistiques, Imagerie, Optimisation Numérique") University of Orléans [link]

Title : "Segmentation interactive d'images médicales par contours interactifs 2D répercutés en 3D" [pdf]
Current status: RD engineer (image processing) at Altatech (Grenoble, France)

2008 : Julien Waechter, Master in Computer Science ("Informatique, spécialité Informatique de l'Image et du Calcul Intensif"), University of Strasbourg

Title: "Intégration du tracking et du recalage d'outils chirurgicaux dans une architecture logicielle à base de composants" [pdf]
Current status: RD engineer at IRCAD

2007 : Mickael Hoarau, french "école d'ingénieur" ENSPS (now TPS), University of Strasbourg

Title : "Segmentation d'images médicales 3D par contours actifs interactifs" [pdf]

2006 : Stéphan Kiss, Master in Computer Science ("Informatique, spécialité Image, Son et Multimédia"), University of Bordeaux

Title : "Débruitage d'images médicales par diffusion anisotrope: un outil interactif de détermination automatique des paramètres" [pdf]

2005 : Benoît Deville, Master in Computer Science, University of Reims

Title : "Création d'un prototype de segmentation interactive 3D par modèles déformables" [pdf]