2022-23

This Master project takes place in the context of an Interdisciplinary seed research project of ITI HealthTech, focused on parsimonious technologies and methods for upper limb rehabilitation. More precisely, the background is the development of robot-assisted therapies for the rehabilitation of neuro-motor function disorders, which have been notably developing in recent years. At the opposite of most existing paradigms, the project emphasizes the need for novel low cost technologies that is a major challenge for the development of robotic assistance for rehabilitation. Besides the development of a robotic structure itself, the perception of the patient muscular activity is important to propose quantitative evaluation of the patient recovery. In this field, existing technologies mostly rely on EMG measurements systems whose limitations in terms of repeatability and selectivity are well known. In the present project, we aim at making proofs of concept using disruptive solutions based on emerging technologies for force sensing.

• Master project supervised by B. Bayle & M. Gandolla
• Master student: Mahrukh Azhar

Attention-deficit/hyperactivity disorder (ADHD) is a neuropsychiatric disorder and patients suffer from inattention, hyperactivity and impulsivity. Since pharmacological treatment may yield adverse cognitive side effects, non-pharmacological neurostimulation is suggested as an alternative. The project investigates the impact of auditory beat stimulation on visual attention in healthy subjects showing ADHD symptoms. The student will perform psychophysical experiments in the presence of binaural beats or isochronic tones and will identify stimulation parameters that improve best visual attention in subjects.

• Master project supervised by A. Hutt & A. Bonnefond
• HealthTech Master student: Gabriel Alves Castro

Interventional radiology is a practice that remains largely less developed than surgery, even though in many cases it can offer a non-surgical alternative to severe conditions. Non-vascular interventional radiology is based on percutaneous procedures that are particularly long and relatively poorly codified. This project aims to provide tools for training in percutaneous procedures, with two main objectives: to formalize the learning of the gesture and to propose an assistance for this learning. The first objective will be to study and formalize the learning and evaluation methods of the learner. The gesture guidance, based on the evaluation, will aim at first providing indications as a mentor would do, and secondly to propose a guidance by effort feedback to accompany the practitioner, whose contribution will be evaluated.

• Master project supervised by H. Omran, B. Bayle & F. Nageotte
• HealthTech Master student: Thomas Cesare

We aim at developing a real-time model for the control of continuous cable-actuated robots, such as the STRAS robot developed at ICube. This model could then be used for automatic or semi-automatic control. Indeed, the originality of our approach lies in the fact that we address the problem of deformation by solving inverse finite elements simulations in real time in the control loop of the robot. This requires accurate modeling of robot’s behavior and interactions with the environment. In addition, to consider robot control, numerical calculations must be performed in a real-time context. In this project we want to implement in SOFA a real-time cable model based on Cosserat's theory. We also want to develop numerical solutions to simulate the constraints with the environment (collision, friction, ...). Moreover, we wish to propose solutions to parameterize these numerical models in order to obtain a predictive simulation of the macroscopic behavior of the robot.

• Master project supervised by B. Rosa & H. Courtecuisse
• HealthTech Master student: Chiara Cignolini

Pancreatic cancer is predicted to become 2nd cause of death by cancer in 10 years from now. To counter this trend, early diagnosis is the only way to significantly improve survival rate, but as standard 3D medical imaging fails to achieve early detection, endoscopic ultrasound (EUS) stands as the only viable option today. This technique is not widely available due its complexity. One challenging aspect of EUS is the complex interpretation of ultrasound images during the examination. It is not rare for non-expert to miss the screening of parts of the pancreas during the procedure, leaving tumors undetected. During this project, we will focus on pancreas part detection in real-time. We aim at developing deep learning methods and deploy them in the operation room conditions to support non-expert clinicians in their practice.

• Master project supervised by N. Padoy, L. Sosa-Valencia & J.-P. Mazellier
• HealthTech Master student: Erik Feragotto

The project aims to build a comprehensive exhaustive classification of combined hepatocellularcholangiocarcinomas (cHCC-CCA) based on their multilevel morphological features and identify prognostic subgroups allowing to propose a tailored management of patients. This will be achieved by the development of new multimodal machine learning methods, based on imagomics that integrate multisource imaging data (histopathological and mass spectrometry imaging).

• Master project supervised by C. Wemmert & V. Paradis
• HealthTech Master student: Elena Ferraguti

Bone remodeling is a lifelong process which involves the removal of old or damaged bone and the consecutive replacement with new bone. Normal bone remodeling guarantees the maintenance of a healthy skeleton to provide support for withstanding daily activities and mechanical loads. Nonetheless, this complex mechanobiological process may be altered by several factors such as pathologies (i.e. cancer, osteoporosis) or prosthesis insertion. In the latter case, damage may occur and it can modify bone remodeling or lead to bone fracture. Our objective is to experimentally identify the mechanical processes and behaviors involved and to couple them with a bone remodeling model. Then, the results will allow us to optimize the prosthesis both mechanically and geometrically.

• Master project supervised by N. Bahlouli, C. Cluzel & R. Allena
• HealthTech Master student: Marie Gesser

This project aims to improve an existing anatomical finite element model of the vestibule of the human inner ear, and to use it to develop the knowledge of the physiology of the balance system in its normal behavior, and its pathological dysfunction. To this end, an extended database of medical images (CT-scans and MRIs) will be analyzed and processed. The goal is to establish the variability of the individual anatomy of the inner ear, and to build automatically a personalized finite element model of a given patient, thus leading to specific diagnostic and therapeutic tools.

• Master project supervised by D. Baumgartner & A. Charpiot
• HealthTech Master student: Ethan Giolito

Human operators can recognize intrahepatic structures using ultrasound images in swine anatomy, even in the first time (domain shift). How they accomplish this remains an open question. Spatial information can be possibly exploited to understand operator’s attention, reducing US videos to key US clips / images. These key US clips can be used to train AI models to identify the structures being truly attended by operator

• Master project supervised by A. Karargyris & J. Verde
• HealthTech Master student: Dylan Meckes

Manually analysing histological slices is the current gold standard in pathology for cancerous tissue diagnosis. However, the overall method is complex, requiring sample preparation and slice screening by operator. This project proposes to take advantage of new preparation-free 3D imaging techniques of living tissue to produce medical input at a significantly higher rate. Nevertheless, the large volume of generated data is only relevant when an automated system supports pathologists in their analysis. To this aim, we will study unsupervised deep learning methods to demonstrate their capability in supporting surgeons for intra-operative and per-operative decisions.

• Master project supervised by T. Lampert, A. Venkatasamy & J.-P. Mazellier
• HealthTech Master student: Laetitia Rebière