Research topics - details

---

Images and models for Video-Assisted Thoracic Surgery

Project Labex CAMI PhD Thesis (2016-2019). Collaboration: TIMC, IMAG, Université Grenoble-Alpes. Staff: Pablo Alvarez (Thesis, tel-03104312), Simon Rouzé MD (Thesis, tel-03694901), Yohan Payan, Matthieu Chabanas (co-supervisors TIMC).

Project ANR TecSan VATSop (2020-). Collaboration: TIMC, IMAG, Université Grenoble-Alpes, MIMESIS, INRIA (Strasbourg). Staff: Pablo Alvarez (Post-Doc, Grenoble), Simon Rouzé MD, Valentin Boussot (Master Thesis, Thesis), Miguel Castro (Ing.), Tang Hui, Chen Qingmei (Master).

The objective is the intraoperative localization of pulmonary nodules in video-assisted thoracoscopy surgery (VATS).The main idea of our approach is to couple and integrate a biomechanical model of the lung to the image processing needed for guidance.


Rouzé S., Alvarez P., Delatour B., Flécher E., Dillenseger J.-L., Verhoye J.-P., “Localisation de nodules pulmonaires en réalité augmentée grâce au Cone Beam Computed Tomography (CBCT) en vidéo-thoracoscopie”, Bulletin de l’Académie Nationale de Médecine, 202, 8-9, 2018, pp. 1897-1908, doi: 10.1016/S0001-4079(19)30183-9.
Alvarez P., Rouzé S., Miga M., Payan Y., Dillenseger J.-L., Chabanas M., “A hybrid, image-based and biomechanics-based registration approach to markerless intraoperative nodule localization during video-assisted thoracoscopic surgery”, Medical Image Analysis, 2021, pp. 101983, doi: 10.1016/j.media.2021.101983

Estimation of lung deformities between dorsal and lateral decubitus positions.

The preoperative volume is acquired in supine CT and the intraoperative volume is acquired in lateral CBCT. The idea is to analyse and estimate the deformations between these two modalities using elastic registration. In a second step, we try to predict these deformations by a statistical model.


Alvarez P., Chabanas M., Rouzé S., Castro M., Payan Y., Dillenseger J.-L., “Lung deformation between preoperative CT and intraoperative CBCT for Thoracoscopic Surgery: a case study”, SPIE Medical Imaging: Image-Guided Procedures, Robotic Interventions, and Modeling, Houston, 2018, pp.S45-S46, doi: 10.1117/12.229393
Alvarez P., Chabanas M., Sikora S., Rouzé S., Payan Y., Dillenseger J.-L., “Measurement and analysis of lobar lung deformation after a change of patient position during video-assisted thoracoscopic surgery”, IEEE Transactions on Biomedical Engineering, 2022, doi: 10.1109/tbme.2022.3205458.
Boussot V., Dillenseger, J.-L., “Modèle statistique pour la prédiction de la déformation du poumon pendant la chirurgie thoracique vidéo-assistée“, RITS 2022, Brest, 2022.
Boussot V., Dillenseger J.-L., “Statistical model for the prediction of lung deformation during video-assisted thoracoscopic surgery”, SPIE Medical Imaging, Image-Guided Procedures, Robotic Interventions, and Modeling, SPIE vol. 12466, San Diego, 2023, doi: 10.1117/12.2646983.

Biomechanical modeling of the deflation

Starting from the previous model, the idea is to develop a biomechanical model which simulates the deflation due to pneumothorax.


Alvarez P., Rouzé S., Chabanas M., Castro M., Payan Y., Dillenseger J.-L., “Modeling lung deflation during video-assisted thoracoscopy surgery for the localization of small nodules", EUROMECH Colloquium 595: Biomechanics and computer assisted surgery meets medical reality, Villeneuve d’Asq, 2017
Alvarez P., Rouzé S., Chabanas M., Castro M., Payan Y., Dillenseger J.-L., “A biomechanical model of lung deflation during VATS for the localization of small nodules”, Surgetica, Strasbourg, 2017.
Alvarez P., Narasimhan S. Rouzé S., Dillenseger J.-L., Payan Y., Miga M., Chabanas M., “Biphasic model of lung deformations for Video-Assisted Thoracoscopic Surgery (VATS)”, IEEE ISBI 2019, Venise, 2019, pp 1367-1371, doi: 10.1109/ISBI.2019.8759219.

---

---

Fall detection using stereoscopic thermal cameras or coupled thermal/depth cameras

Project ANR TecSan PRuDENCE (2016-2019). Collaboration : NeoTec Vision (Pacé), ECAM Rennes - Louis de Broglie (Bruz), Living Lab Activ Ageing UTT (Troyes), équipe accueil EA 2694 (Lille). Staff : Yannick Zoetgnandé (Thesis, tel-03118117), Imen Halima (Thesis ECAM Rennes, tel-03212630), Soumaya Msaad (Thesis supervised by Guy Carrault, tel-03815836), Abderamane Abakar Bechir (Master), Mathis Fleury, Cédric Moubri-Tournès, OscarTanguille (ESIR training period).

One of the goals of this project is the fall detection of elderly people by either a pair of stereoscopic low resolution thermal cameras or the coupling of a depth and a thermal camera.


Lan Hing Ting K., Voilmy D., Dessinger G., Gauthier V., Dillenseger J-L., Carrault G., Laferté J-M., Fougères A-J., “Co-designing a falls detection device: combining concerns for human motion and elders needs”, Workshop Visual user interfaces for human motion, ACM AVI 2020, 2020.
Lan Hing Ting K., Voilmy D., Dessinger G., Gauthier V., Dillenseger J-L., Carrault G., Laferté J-M., Fougères A-J., “Causes et conséquences de la chute : les comprendre pour informer la conception d’un dispositif de prévention”, Colloque francophone sur la chute de la personne âgée, Valenciennes, 2021.

People tracking and activity recognition by coupling a depth and a thermal camera

The tracking of people is performed on a pair of images (depth sensor / thermal image) by a particle filter approach or by deep learning (Imen Halima).


Halima I., Laferté J.-M., Dillenseger J.-L., Cormier G., Fougères A.-J., "Sensors fusion for head tracking using Particle filter in a context of falls detection”, 1st International Conference on Advances in Signal Processing and Artificial Intelligence (ASPAI' 2019), Barcelona, 2019,pp. 134-139.
Halima I., Laferté J.-M., Cormier G., Fougères A.-J., Dillenseger J.-L., “ Depth and thermal information fusion for head tracking using particle filter in a fall detection context”, Integrated Computer-Aided Engineering, 27, 2, 2020, pp. 195-208, doi: 10.3233/ICA-190615.

We addressed activity tracking by deep learning. For this we created a learning base of activities on 3 different sites (an empty room, an apartment in a living lab and a real apartment) with 3 different people. We trained an SSD network to estimate 4 postures (sitting, standing, lying on the bed and lying on the floor) (Imen Halima).

Fall detection of elderly people by either a pair of stereoscopic low resolution thermal cameras

Thermal cameras stereo calibration and subpixel reconstruction

The chosen thermal sensors (Lepton 2 FLIR) are low price but also has a very low resolution (80x60 pixels). We built a car-bulbs calibration grid adapted to thermal imaging cameras (Mathis Fleury, Cédric Moubri-Tournès, Oscar Tanguille). We have also developed a robust low resolution stereo-camera calibration framework (Yannick Zoetgnandé).
The image content and low resolution of the sensors have also an impact on the reconstruction of the scene by stereoscopic vision. We have developed a reconstruction protocol based on the robust extraction of visual features by phase congruency, feature matching and sub-pixel precision disparity estimation by phase coherence (Yannick Zoetgnandé).


Zoetgnandé Y., Fougères A.-J., Cormier G., Dillenseger J.-L., “Robust low resolution thermal stereo camera calibration”, 11th International Conference on Machine Vision (ICMV 2018), proc. SPIE 11041, Munich, 2018, pp. 11041-1D, doi: 10.1117/12.2523440 .
Zoetgnandé Y., Fougères A.-J., Cormier G., Dillenseger J.-L., “Sub-pixel matching method for low resolution thermal stereo images”, Infrared Physics and Technology, 105, 2020, pp. 103161, doi: 10.1016/j.infrared.2019.103161.

Thermal Images super-resolution

The goal is the improvment of the thermal images resolution by a ratio 3 or 4. we develloped a CNN-based approach with edge salliency (Yannick Zoetgnandé during an exchange at Ryerson University, Toronto, Canada, MITACS Globalink Research Award FR26353, collaboration with Pr Alirezaie,).


Zoetgnandé Y., Dillenseger J.-L., Alirezaie J., “Edge focused super-resolution of thermal images”, International Joint Conference on Neural Networks (IJCNN), Budapest, 2019, doi: 10.1109/IJCNN.2019.8852320.

Fall detection by ground points learning

Despite the sub-pixel accuracy, the classical stereoscopic reconstruction did not allow an accurate reconstruction of the scene, in particular the ground plane and the person (0.5 m of uncertainty in depth). We therefore opted for a solution without calibration, where a system learns to classify points in space as belonging to "ground/not ground" (either by SVM or by deep learning). Then, from a stereo-thermal image pair, the device extracts and matches visual features of the person and estimates whether these features are on the ground or not. This allows to detect or not a fall (Yannick Zoetgnandé).

Zoetgnandé Y., Dillenseger J.-L., “A generic interpretable fall detection framework based on low-resolution thermal images”, Journées de la Recherche en Informatique au Burkina Faso (JRI-2021), Bobo Dioulasso, 2021, 13 pages, doi: 10.4108/eai.11-11-2021.2317972.

Detection of activities by transfer learning

We also tried to estimate the activity of elderly people (walking, crawling, sitting and falling, lying down and falling, etc.) from a single low resolution thermal camera.As we did not have enough data for training, we explored several solutions where the training of a network performed on anotated RGB data was used to process the thermal data from our cameras


Zoetgnandé Y., Dillenseger J.-L., “Domain generalization for activity recognition: Learn from visible, infer with thermal”, 11 th International Conference on Pattern Recognition Applications and Methods, 2022, pp.722-729, doi: 10.5220/0010906300003122.

Prevention of falls and estimation of the frailty of elderly people by monitoring activities.

The objective is to recover the activities of the elderly person recognized by the previous techniques and to make a longitudinal follow-up and to extract statistics aiming at measuring the degree of frailty of the person (Soumaya Msaad).

Msaad S., Dillenseger J.-L., Cormier G., Carrault G., “Detection of changes in the behaviour of the elderly person”, IEEE-EMBS, Guadalajara, 2021, pp. 6995-6998, doi: 10.1109/embc46164.2021.9630971.
Msaad S., Dillenseger J.-L., Carrault G., “Interest of the minimum edit distance to detect behaviour change of the elderly person”, IEEE-EMBS, Guadalajara, 2021, pp. 7377-7380, doi: 10.1109/embc46164.2021.9629665.
Msaad S., Zoetgnandé Y., Dillenseger J.-L., Carrault G., “Detecting change in the routine of the elderly”, Measurement: Sensors, 2022, pp. 100418, doi: 10.1016/j.measen.2022.100418.

---

---

Cardiac arrhythmias treatment

Grant: ANR TecSan CardioUSgHIFU (2012-2015). Collaboration: Institut Langevin, ESPCI (Paris), LabTau, INSERM U 1032 (Lyon), Vermont (Tours). Staff: Zulma Sandoval (Thesis, tel-01241529) , Pham Chi Hieu (Master).

Grant: ANR TecSan Chorus (2018-). Collaboration: LabTau, INSERM U 1032 (Lyon), IHU Lyric (Bordeaux) Vermont (Tours). Staff: Batoul Dahman (Master and Thesis, tel-03892658)

This project is focuses on the isolation of the pulmonary veins for the treatment of cardiac fibrillation using HIFU. Our goal is the transesophageal therapy guiding on ultrasound images.

Therapy Guidance using 2D transesophageal US/preoperative volume registration

Similarity metric evaluation on synthetic data.

This work is to evaluate the several intensity based similarity measures used in the registration of ultrasound and CT images of the left atrium and the pulmonary veins. The evaluation is performed on numerical phantoms.

Visible Human slice. Segmented slice. CT scan simulation. Transoesophageal US image simulation.

Sandoval Z., Dillenseger J.-L., “Intensity-based Similarity Measures Evaluation for CT to Ultrasound 2D Registration”, IRBM, 34, 4-5, 2013, pp. 278-282.
Sandoval Z., Dillenseger J.-L., “Evaluation of Computed Tomography to Ultrasound 2D Image Registration for Atrial Fibrillation Treatment”, Computing in Cardiology, Zaragoza, 2013, pp. 245-248.

2D transoesophageal US/preoperative volume registration

ANRCardioUSgHIFU :
The main idea is to use the esophagus as anatomical constraint in order to reduce degrees of freedom.

Esophagus etraction; esophagus-based volume ; US 2D/ CT 2D registration; 2D transoesophageal US/preoperative volume registration

Sandoval Z., Castro M., Alirezaie J., Lafon C., Bessière F., Dillenseger J.-L., “Transesophageal 2D ultrasound to 3D computed tomography registration for the guidance of a cardiac arrhythmia therapy”, Physics in Medicine and Biology, 63, 15, 2018, pp. 155007, doi 10.1088/1361-6560/aad29a.

ANR Chorus :
One of the objetvives of the ANR Chorus project is to develop a dual mode transesophageal probe with a 2 perpendicular US imaging planes. The idea is now to relax the previous anatomical constraints and to refine the registration using this new information. Classical approaches as well as deep learning ones (2D/2D and 2D/3D rigid registration with supervised  learning and 2D/2D elsatic registration with unservised learning) have been studied.

Biplan US / CT registration

Dahman B., Dillenseger, J.-L., “High Intensity Focused Ultrasound (HIFU) Therapy Guidance System by Image-based Registration for Patients with Cardiac Fibrillation”, Computing in Cardiology, Singapore, 2019, doi: 10.22489/CinC.2019.315.
Dahman B., Dillenseger, J.-L.,“ Transesophogeal HIFU cardiac fibrilation therapy guidance by 2 two perpendicular US images", Surgetica 2019, Rennes, 2019.
Dahman B., Dillenseger J.-L., “Deformable US/CT Image Registration with a Convolutional Neural Network for Cardiac Arrhythmia Therapy”, IEEE-EMBS, Montreal, 2020, pp. 2011-2014, doi: 10.1109/embc44109.2020.9175386.
Dahman B., Bessière F., Dillenseger J.-L., “Ultrasound to CT rigid image registration using CNN for the HIFU treatment of heart arrhythmias”, SPIE Medical Imaging, Image-Guided Procedures, Robotic Interventions, and Modeling, SPIE vol 12034, San Diego, 2022, pp. 246-252, doi: 10.1117/12.2612348.

Left Atrium segmentation

The left atrium and pulmonary veins segmentation in pre-operative CT is performed by a Multi-Atlas-Based approach refined by local region gowing.

Multi-atlas segmentation framework.

Tobon-Gomez C., Geers A. J., Peters J., Weese J., Pinto K., Karim R., Ammar M., Daoudi A., Margeta J., Sandoval Z., Stender B., Zheng Y., Zuluaga M., Betancur J., Ayache N., Chikh M. A., Dillenseger J.-L., Kelm, M., Mahmoudi S., Ourselin S., Schaeffer T., Schlaefer A., Ravazi R., Rhode K. S., “Benchmark for algorithms segmenting the left atrium from 3D CT and MRI datasets”,IEEE transactions on Medical Imaging, 34, 7, 2015, pp. 1460-1473
Sandoval Z., Betancur, J., Dillenseger J.-L., “Multi-Atlas-Based Segmentation of the Left Atrium and Pulmonary Veins”, Statistical Atlases and Computational Models of the Heart - MICCAI 2013, Nagoya, 2013.
Sandoval Z., Dillenseger J.-L., “Thorax tissues segmentation: a first step for a dynamic beating heart digital phantom”, RITS 2015, Dourdan, pp. 158-159

---

---

Augmented reality in robotic surgery of tumors

Project: Labex CAMI PhD Thesis (2015-2018). Collaboration: Laboratoire d’Informatique, de Robotique et de Microélectronique de Montpellier (LIRMM) . Staff: Shen Jun (Thesis, tel-02570502), Philippe Poignet et Nabil Zemiti (co-supervisors LIRMM).

The goal is to propose an Augmented Reality (AR)-based framework for the guidance of transoral robotic surgery of base tongue tumor.

Augmented Reality of 3D Ultrasound information embeded to the binocular robotic vision

The Augmented Reality requires the 3D calibration and localization of the 3D ultrasound imagee (a new calibration phantom with a single 3D marker has been proposed) and the stereoscopic calibration and localization of the stereoscopic camera. Tests on physical phantoms (insertion of needles and target cuts) were conducted to validate the AR framework.

Augmented reality experience: a) Calibration and localization of 3D ultrasound, b) MRI acquisition, c) Segmentation of ultrasound and d) delineation of the target in augmented reality

Shen J., Zemiti N., Caravaca, O., Simon A., Dillenseger J.-L., Poignet P., “ Intraoperative 3D ultrasound probe calibration", RITS 2017, Lyon
Shen J., Courtin A., Viquesne, A., Zemiti N., Caravaca, O., Simon A., Dillenseger J.-L., Poignet P., “Augmented perception in transoral robotic surgery for tongue base cancer”, CRAS Workshop, Montpellier, 2017.
Shen J., Zemiti N., Caravaca, O., Simon A., Dillenseger J.-L., Poignet P., “Augmented reality visualization based on 3D ultrasonography”, Surgetica, Strasbourg, 2017.
Shen J., Zemiti N., Dillenseger J.-L., Poignet P., “Fast and simple automatic 3D ultrasound probe calibration based on 3D printed phantom and an untracked marker”, IEEE-EMBC, Honolulu, 2018, pp. 878-882.
Shen J., Zemiti N., Taoum C., Aiche G., Dillenseger J.-L., Rouanet P., Poignet P., "Transrectal ultrasound image-based real-time augmented reality guidance in robot-assisted laparoscopic rectal surgery: a proof-of-concept study", International Journal of Computer Assisted Radiology and Surgery, 15, 3, 2020, pp. 531–543, doi: 10.1007/s11548-019-02100-2

---

---

Medical image characterization using Mixture Models

Collaboration: CRIBS, LIST, SouthEast University, Nanjing. Staff: Tang Hui (co-supervisor) Bi Hui (Thesis)

Gaussian Mixture Model incorporating spatial information

Gaussian Mixture Model weighted by boundary information. Integration of the neighborhood information in the Gaussian Mixture parameters estimation process.

Kidney image characterization by Gaussian Mixture a) Without neighborhood information. b) Integrating neighborhood information

Tang H., Dillenseger J.-L., Luo L. M., "A vectorial image classification method based on neighborhood weighted Gaussian Mixture model", IEEE-EMBC 2008, Vancouver, 2008, pp. 1922-1925.
Tang H., Dillenseger J.-L., Bao X. D., Luo L. M., " A vectorial image soft segmentation method based on neighborhood weighted Gaussian mixture model", Computerized Medical Imaging and Graphics, 33, 8, 2009, pp. 644-650.

aussian Mixture Model and saliency

Gaussian Mixture Model guided by image content-based information

Extraction of the saliency image, integration of this information to guide the estimation of the Gaussian Mixture

Bi H., Tang H., Yang, G.Y., Li, B.S., Shu, H.Z., Dillenseger J.-L., “Accurate Image Segmentation Using Gaussian Mixture Model with Saliency Map”, Pattern Analysis and Applications, 27,3, 2018, pp. 869-878, doi 10.1007/s10044-017-0672-1.

Mixture of Rayleigh distributions Mixture Model weighted by neighborhood information. Ultrasound image characterization.

Characterization of US speckle using a Rayleigh distributions Mixture Model weighted by neighborhood information

Characterization of an ultrasound image of the abdomen (simulation) Rayleigh Mixture a) without and b) with neighborhood information

Bi H., Tang H., Shu H.Z., Dillenseger J.-L., “Bounded Rayleigh mixture model for ultrasound image segmentation”, 8th International Conference on Graphic and Image Processing, Tokyo, 2016, pp. 47.1-47.5. Winner of best presentation session 4.
Bi H., Tang H., Yang, G.Y., Li, B.S., Shu, H.Z., Dillenseger J.-L., “Fast Segmentation of Ultrasound Images by Incorporating Spatial Information into Rayleigh Mixture Model”, IET Image Processing,11, 12, 2017, doi 10.1049/iet-ipr.2017.0166, pp. 1188-1196.
Bi H., Yang G.Y., Tang H., Jiang Y.B., Shu H.Z., Dillenseger J.-L., “Fast and accurate segmentation method of Active Shape Model with Rayleigh Mixture Model clustering for prostate ultrasound images”, acceptet at Computers Methods and Programs in Biomedicine, 2019, doi: 10.1016/j.cmpb.2019.105097.

---

---

Biometry for the diagnosis of congenital heart defects

Collaboration: Jean-Marc Schleich (PH, CHU Rennes), Tariq Abdulla (Univ. Loughborough), Lucile Houyel (Centre Chirurgical Marie Lannelongue, Le Plessis-Robinson) .

The goal is to propose a tool for the classification and the grading of cardiocongenital heart defects from some heart structures relative position and orientation ( aortic valve and pulmonary valve annuli, main coronary arteries orifices, etc.).

Ex vitro biometry of the heart : a) anatomical landmaks 2D localization tool, b) 3D localization tool, c) geometrical modelling of some cardiac structures.

Abdulla T., Dillenseger J.-L., Summers R., Paul J.-F., Houyel L., Schleich J.-M., " I point my heart with the tip of my fingers - Biometry for the diagnosis congenital heart defects", IRBM, 34, 4-5, 2013, pp. 311-314.

Previous project: Visualization in cardiac embriology

Three-dimensional reconstruction and visualization of human embryonic heart from 8 to 13 weeks of gestational age.

>

Schleich J.-M., Dillenseger J.-L., Loeuillet L., Moulinoux J.-P., Almange C., "Three-dimensional reconstruction and morphological measurements of human embryonic hearts. A new diagnostic and quantitative method applicable to fetuses younger than 13 weeks of gestation", Pediatric and Developmental Pathology, 8(4), 2005, pp. 463-473.

This work has been performed in collaboration with Dr Jean-Marc Schleich who works on a new dynamic 3-D virtual image method for teaching normal heart development.

Introduction to the atrial septation animation (video)

Schleich J.-M., Dillenseger J.-L., Andru S., Coatrieux J.-L., Almange C., "Understanding Normal Cardiac Development using Animated Models ", IEEE Computer Graphics & Applications, 22, 1, 2002, pp. 14-19.
Schleich J.-M., Dillenseger J.-L., Houyel L., Almange C., Anderson R. H., "A new dynamic 3-D virtual image method for teaching normal human atrial septation", Anatomical Sciences Education, 2, 2, 2009, pp. 69-77.

---

---

HIFU therapy for the focal prostate tumor treatment

Grant: ANR TecSan MULTIP (2008-2012). Collaboration: Edap (Vaulx en Vellin), LabTau, INSERM U 1032 (Lyon), Imasonic (Voray/l'Ognon). Staff: Carole Garnier (post doc) Ke Wu (Thesis, tel-00962028)

ANR TecSan MULTIP Project (Matrice de transducteur Ultrasonore pour La Thérapie et l’Imagerie de la Prostate). This project is aimed to propose a preoperative High Intensity Focused Ultrasound therapy for the focal prostate tumor treatment.

MRI/US prostate registration

The main topic is here to transfer the anatomical prostate information (central zone, ...) extracted during the preoperative planning from T2 MRI on the preoperative ultrasound image. This should allow retrieving the tumor location during the therapy.
Two strategies have been explored: surface to surface registration after a surface segementation steps on both modalities and volume to volume aproach after the characterization of the volumic information by texture analysis.

Surface to surface registration

The extraction of the prostate surface on ultrasound volulme has been performed during Carole Garnier's PhD Thesis. Optimal Surface detection (OSD) and Discrete Dynamic Contour (DDC) methods have been explored. The OSD method was adapted for joint extraction of the bladder, colon and prostate Optimal Surface Detection (OSD) on T2 MRI (Wu Ke).

a) Prostate surface mesh extracted on a 3D US volume using Optimal Surface Detection. b) Prostate (green) and bladder (red) surface meshes extracted on a T2 MRI volume using multi-objects Optimal Surface Detection.

Garnier C., Wu K., Dillenseger J.-L., "Bladder segmentation in MRI images using active region growing model", IEEE-EMBC 2011, Boston, 2011, pp. 5702-5705.
Wu K, Garnier C., Dillenseger J.-L., “Prostate segmentation on T2 MRI using Optimal Surface Detection", IRBM, 34, 4-5, 2013, pp. 287-290.
Wu K., Garnier C., Alirezaie J., Dillenseger J.-L., ”Adaptation and evaluation of the multiple organs OSD for T2 MRI prostate segmentation”, EMBS 2014, Chicago, pp. 4687-4690.

Volume characterization by texture analysis

The characterization of the prostatic volume in ultrasound images is the subject of Wu Ke's PhD Thesis. The speckle is considered as a texture which is characterized using moment invariants. We found that orthogonal moment invariants were able to characterize not only the several region on US images but also the boundary between two regions. This property has been used in the context of Graph Cut segmentation.

a) Original prostate image b) orthogonal moment invariant feature characterizing the regions c) orthogonal moment invariant feature characterizing the contours.

Wu K., Shu H.Z., Dillenseger J.-L., “Region and boundary feature estimation on ultrasound images using moment invariants”, Computers Methods and Programs in Biomedicine, 113, 2, 2014, pp. 446-455.

---

---

HCC interstitial HIFU treatment

Grant: RNTS SUTI (2008-2012). Collaboration : LabTau, INSERM U 1032 (Lyon), FEMTO-LPMO CNRS UMR6174, (Besançon), Edap (Vaulx en Vellin), Imasonic (Voray/l'Ognon), Theraclion (Paris) IGT (Pessac). Staff: Carole Garnier (Master), Simeon Esneault (Thesis, tel-00497749), Zheng Yang (Master).

RNTS SUTI project (Sondes Ultrasonores pour la Thérapie et l'Imagerie). LTSI is in charge to propose and to develop new preoperative planning methods. This planning can be divided into two aspects: the placement of the ultrasound probe into the patient's anatomy and the previsionof the therapy effect according to the ultrasound probe parameters.

Treatment planning and guidance assistance

Characterization of the anatomy (Liver, tumor and vascularization)

Liver and hepatic vascularization segmentation using a Min-cut/Max-flow algorithm (S. Esneault, T. Pham, K. Torres)

a) Local vessel model adjustment (identified by geometrical moments) withion the graph cut. b) Tumor and surrounding vascularization.

Esneault S., Hraiech N., Delabrousse E., Dillenseger J.-L., "Graph cut liver segmentation for interstitial ultrasound therapy", IEEE-EMBC 2007, Lyon, 2007, pp. 5247-5250.
Esneault S., Lafon C., Dillenseger J.-L., " Liver vessels segmentation using a hybrid geometrical moments/graph cuts method", IEEE transactions on Biomedical Engineering, 57, 2, 2010, pp. 276-283.

Therapy lesion planning and control

3D modelling of the effect of an ultrasound probe (C. Garnier, S. Esneault). The pressure field is estimated using Rayleigh integral. The parameter madifications acording to the temperature are taken into account. The bioheat equation is solved using finite differences and algebraically after a Fourier transformation over the space coordinates. Optimal phase estimation for dynamic focussing (Z.H. Yang).

Simulation of the pressure field delivered by the probe, the tissue temperature and the induced necrosis

Garnier G., Lafon C., Dillenseger J.-L., "3D modeling of the thermal coagulation necrosis induced by an interstitial ultrasonic transducer", IEEE transactions on Biomedical Engineering, 55, 2, 2008, pp. 833-837.
Dillenseger J.-L., Garnier G., " Acoustical power computation acceleration techniques for the planning of ultrasound therapy", ISBI'08, Paris, 2008, pp. 1203-1206.
Dillenseger J.-L., Esneault S., " Fast FFT-based bioheat transfer equation computation ", Computers in Biology and Medicine, 40, 2, 2010, pp. 119-123.
Yang Z.H., Dillenseger J.-L., "Phase estimation for a phased array therapeutic interstitial ultrasound probe", IEEE-EMBC 2012, San Diego, 2012, pp. 472-475.

---

---

Image guided surgery in urology

Collaboration : J-J Patard (PU-PH CHU Rennes). Staff: Hélène Guillaume (Thesis), Hui Tang (Thesis, tel-00355629), Philippe Rolland (UTC), Soizic Laguitton (Master), Hamza Khene (Master)

Patient specific anatomical framework

Intra-subject uroscans matching

Registration/modeling spherical harmonics based technique applied on partial information (H. Guillaume)

Dillenseger J.-L., Guillaume H., Patard J.-J., " Spherical harmonics based intra subject 3D kidney modeling/registration technique applied on partial information", IEEE transactions on Biomedical Engineering, 53, 11, 2006, pp. 2185-2193.

3D/3D kidney registration using local Mutual Information maximization (H. Tang)

Tang H., Dillenseger J.-L., Luo L. M., "Intra subject 3D/3D kidney registration using local Mutual Information maximization", IEEE-EMBC 2007, Lyon, 2007, pp. 6379-6382.

Renal volume segmentation/labeling

Vectorial image classification method based on neighborhood weighted Gaussian Mixture model (H. Tang)

Tang H., Dillenseger J.-L., Luo L. M., "A vectorial image classification method based on neighborhood weighted Gaussian Mixture model", IEEE-EMBC 2008, Vancouver, 2008, pp. 1922-1925.
Tang H., Dillenseger J.-L., Bao X. D., Luo L. M., " A vectorial image soft segmentation method based on neighborhood weighted Gaussian mixture model", Computerized Medical Imaging and Graphics, 33, 8, 2009, pp. 644-650.

Renal volume visualisation

Volume rendering technique on data described by weighted Gaussian Mixture model (H. Tang).

Tang H., Dillenseger J.-L., Bao, X. D., Luo, L. M. , "A multi-volume visualization method for spatial aligned volumes after 3D/3D image registration", IRBM, 32, 3, 2011, pp. 195-203.

Preoperative planning

Visual optimal trajectory definition of a percutaneous tract for nephrolithotomy or radiofrequency tumor treatment (P. Rolland).

Dillenseger J.-L., Rolland P., Patard J.-J., "A Visual Computer Tool for Percutaneous Nephrolithotomy Preoperative Planning", IEEE-EMBC 2003, Cancun, 2003, pp. 1141-1144.

Simulation of the intervention

Fast simulation of ultrasound images from a CT volume (S. Laguitton)

Laguitton S., Patard J.-J., Dillenseger J.-L., "Fast simulation of ultrasound images from a CT volume", EMBEC’05, Prague, Nov. 2005, papier 1433, 4 pages.
Dillenseger J.-L., Laguitton S., Delabrousse E, "Fast Simulation of Ultrasound Images from a CT volume", Computers in Biology and Medicine, 39, 2, 2009, pp. 180-186.

Intraoperative treatment

Preoperative/Intraoperative registration

2D/3D matching et and tract tracking (H. Khene)

---

---

Visualization for clinical epilepsy studies

Development of complementary visual tools allowing a progressive analysis of the structures and mechanisms involved during a seizure.

Rocha C., Dillenseger J.-L., Coatrieux J.-L., "Multi-Array EEG Signals Mapped with Three Dimensional Images for Clinical Epilepsy Studies", Lecture Notes in Computer Science 1131, Visualization in Biomedical Computing, Höhne and Kikinis Eds., Springer Verlag, 1996, pp. 467-476.

>

---

---

Scientific and generic medical visualization

Visualization tool formalism

Visualization tool basic frame. Collaboration : Beatriz Souza santos (Univ Aveiro, Portugal).

Principle - segmentation during visualization- noise canceling during visualization

Multi-Agent frame. Automatic prepositionning from facial feature labeling. Cortex segmentation using knowledge on the acquisition modality and the morpohological properties