Portrait de Flavie Lavoie-Cardinal

Flavie Lavoie-Cardinal

Membre académique associé
Professeur agrégé, Université Laval, Département de Psychiatrie et de neurosciences
Sujets de recherche
Apprentissage profond
Biologie computationnelle
Biophysique
Détection d'anomalies
IA appliquée
IA en santé
IA pour la science
Modèles de fondation
Modèles génératifs
NeuroIA
Neurosciences
Neurosciences computationnelles
Segmentation d'images médicales
Vision par ordinateur

Biographie

La Dre Lavoie-Cardinal est professeure agrégée au Département de psychiatrie et de neurosciences de l’Université Laval, à Québec, ainsi que titulaire d’une Chaire de recherche du Canada (niveau 2) du CRSNG en nanoscopie intelligente de la plasticité cellulaire. Elle est également chercheuse principale au Centre de recherche CERVO et responsable de l’axe Santé et sciences de la vie à l’Institut intelligence et données (IID) de Québec. Elle est corésponsable de l’axe Perception et contrôle du regroupement stratégique Neuro-IA UNIQUE, ainsi que l’une des corésponsables du regroupement Neuro-IA 1 à l’IVADO.

Elle a obtenu son doctorat en chimie en 2011, suivi de deux stages postdoctoraux, dont l’un dans le groupe du Pr Dr Dr Stefan Hell (lauréat du prix Nobel de chimie 2014), portant sur le développement de techniques de microscopie à super-résolution.

Elle dirige actuellement un groupe de recherche composé de 14 étudiants aux cycles supérieurs et au premier cycle, travaillant à l’interface de la microscopie optique, des neurosciences et de l’intelligence artificielle (IA). Son programme de recherche transdisciplinaire vise à développer de nouvelles stratégies de bio-imagerie assistées par l’IA pour révéler la signature moléculaire de la plasticité synaptique altérée, menant à la neurodégénérescence ou aux troubles cognitifs.

Publications

Chronic Stress Exposure Alters the Gut Barrier: Sex-Specific Effects on Microbiota and Jejunum Tight Junctions
Ellen Doney
Laurence Dion-Albert
Francois Coulombe-Rozon
Natasha Osborne
Renaud Bernatchez
Sam E.J. Paton
Fernanda Neutzling Kaufmann
Roseline Olory Agomma
José L. Solano
Raphael Gaumond
Katarzyna A. Dudek
Joanna Kasia Szyszkowicz
Manon Lebel
Alain Doyen
Marie-Claude Audet
Caroline Menard
Filtering Pixel Latent Variables for Unmixing Volumetric Images
Measurements of different overlapping components require robust unmixing algorithms to convert the raw multi-dimensional measurements to use… (voir plus)ful unmixed images. Such algorithms perform reliable separation of the components when the raw signal is fully resolved and contains enough information to fit curves on the raw distributions. In experimental physics, measurements are often noisy, undersam-pled, or unresolved spatially or spectrally. We propose a novel method where bandpass filters are applied to the latent space of a multi-dimensional convolutional neural network to separate the overlapping signal components and extract each of their relative contributions. Simultaneously processing all dimensions with multi-dimensional convolution kernels empowers the network to combine the information from adjacent pixels and time-or spectral-bins, facilitating component separation in instances where individual pixels lack well-resolved information. We demonstrate the applicability of the method to real experimental physics problems using fluorescence lifetime microscopy and mode decomposition in optical fibers as test cases. The successful application of our approach to these two distinct experimental cases, characterized by different measured distributions, highlights the versatility of our approach in addressing a wide array of imaging tasks.
Filtering Pixel Latent Variables for Unmixing Volumetric Images
Measurements of different overlapping components require robust unmixing algorithms to convert the raw multi-dimensional measurements to use… (voir plus)ful unmixed images. Such algorithms perform reliable separation of the components when the raw signal is fully resolved and contains enough information to fit curves on the raw distributions. In experimental physics, measurements are often noisy, undersam-pled, or unresolved spatially or spectrally. We propose a novel method where bandpass filters are applied to the latent space of a multi-dimensional convolutional neural network to separate the overlapping signal components and extract each of their relative contributions. Simultaneously processing all dimensions with multi-dimensional convolution kernels empowers the network to combine the information from adjacent pixels and time-or spectral-bins, facilitating component separation in instances where individual pixels lack well-resolved information. We demonstrate the applicability of the method to real experimental physics problems using fluorescence lifetime microscopy and mode decomposition in optical fibers as test cases. The successful application of our approach to these two distinct experimental cases, characterized by different measured distributions, highlights the versatility of our approach in addressing a wide array of imaging tasks.
Annotation Cost-Sensitive Deep Active Learning with Limited Data (Student Abstract)
Contextual bandit optimization of super-resolution microscopy
Anthony Bilodeau
Renaud Bernatchez
Albert Michaud-Gagnon
Microscopy analysis neural network to solve detection, enumeration and segmentation from image-level annotations
Anthony Bilodeau
Constantin V. L. Delmas
M. Parent
Paul De Koninck
Microscopy analysis neural network to solve detection, enumeration and segmentation from image-level annotations
Anthony Bilodeau
Constantin V.L. Delmas
Martin Parent
Paul De Koninck
Neuronal activity remodels the F-actin based submembrane lattice in dendrites but not axons of hippocampal neurons
Anthony Bilodeau
Mado Lemieux
Marc-André Gardner
Theresa Wiesner
Gabrielle Laramée
Paul De Koninck