Harmonisation de données d’imagerie de diffusion par auto-encodage adversariel avec respect de contraintes anatomiques

L’acquisition de données d’imagerie de diffusion est suceptible à un important problème de variabilité à travers les différents sites d’acquisition (e.g. Sherbrooke vs. Montréal) et à travers les fabricants (e.g. Siemens vs Philips). Puisque ces données de diffusion ne sont pas invariantes aux protocoles et scanners utilisés, il s’agit d’un problème de taille pour les études cliniques qui désirent utiliser des données provenant de multiples sites d’acquisition.

Detection of human presence integrated into a manipulator robot

To be safe, collaborative robots must be able to detect human presence in their workspace. Different technologies exist to detect intrusions within a predetermined perimeter (physical barriers with switches, light curtains, etc.). However, these technologies require considerable modifications to the workspace, thus hampering their integration. We propose to integrate this functionality into a collaborative robot manipulator in order to minimize the necessary modifications to the workspace and thus reduce the costs of its integration.

Combined cooling, heat and power (CCHP) production from waste heat in the zinc industry

Zinc oxide production by the French process is a high energy-consuming industrial process. The process generates a large amount of waste heat through the melting and vaporization of zinc. Waste heat recovery is an effective way to increase energy efficiency that leads to reduce energy costs and utility consumption in the system.it is important to recover the maximum amount of heat in the process and use it in another application instead of releasing it into the air.

Development of a programmable and highly specific anti-bacterial synthetic live biotherapeutics

Antibiotic resistant bacteria are one of the major threats of the 21st century. These superbugs, which are resistant to virtually all our antibiotic arsenal, can turn benign infections into life threatening diseases. It is estimated that, by 2050, the death toll associated with antibiotic resistant infection will rise to 10 million people. We thus need to act now and develop alternative technologies that can replace conventional antibiotics in order to prevent this crisis.

This project proposes a totally new approach to eliminate antibiotic resistant bacteria.

Anatomical consistency and confidence estimation of cardiac segmentation of ultrasound images using variational auto-encoders

Artificial intelligence shows great promise in the field of medicine. Indeed, neural networks can learn to perform many tasks that would otherwise take hours for physicians to accomplish. For example, neural networks can learn to classify each pixel in cardiac echography images with respect to the anatomical region (cardiac echography segmentation). This allows for faster diagnostic of various pathologies. However, neural networks are not infallible, and it is not trivial to identify these failure cases without expert intervention.

Optimisation des ressources informatiques pour systèmes intelligents distribués: Une approche décisionnelle multi-agents basée sur un modèle microéconomique d’échange en équilibre général.

Le vieillissement de la population représente un défi humanitaire important. Cette transition nous amène à considérer l'utilisation de robots afin d'alléger les besoins en prestation de soins, une solution toutefois contrainte par les limites actuelles de l'informatique. À cet effet, les chercheurs s'inspirent de la capacité cognitive des êtres humains à moduler les stimuli.

Development and implementation of an analytical noise model for axial swept fans

Fan noise is a major environmental issue as it is found in daily appliances such as kitchen fans, hair dryers or computers, in cars such as engine cooling fan systems and Heat and Ventilation Air Conditioning (HVAC) blowers (creating a nuisance not only for passengers but also for pedestrians in city traffic jams), and in all ventilation systems for buildings or tunnels. It possibly causes significant hearing losses and health issues (increased heart diseases), and as such must be mitigated. Blade sweep is one way of reducing noise and thus must be considered in a design cycle.

Utilisation de nuage de points lidar à haute densité en foresterie de précision

Dans l’est du Canada, la gestion durable des forêts de feuillus et des peuplements de bois mixtes passe, entre autres, par l’implémentation d’une sylviculture de précision. En effet, dans ce type de peuplement, l’identification des tiges de haute valeur marchande et de celles à conserver est un élément primordial tant au niveau économique qu’au niveau écologique. Bien que les compagnies forestières bénéficient d’un soutien des technologies de télédétection (ex.

Étude du comportement de rotors co-axiaux

Les appareils à rotors coaxiaux utilisent deux rotors contrarotatifs partageant le même axe de rotation. Pour ces appareils, l’interaction entre les deux rotors est importante pour leurs performances. Ainsi, ce projet porte sur la caractérisation expérimentale de l’effet de l’espacement entre les 2 rotors sur le comportement aérodynamique des drones.

Advanced interconnects manufacturing toolbox for the Integrated Innovation Chain for Digital Prosperity

The ever-increasing complexity of integrated circuits and systems, induced by ubiquitous portable electronics, internet-power applications and artificial intelligence (AI), necessitate to increase the number interconnections between the chips in a package level like never before. These advanced packaging approaches in turn require novel patterning technology to define the connecting lines between the microelectronics chips.

Pages