A5017608284- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Domain Adaptation and Few-Shot Learning
- Advanced Neuroimaging Techniques and Applications
- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- Medical Imaging and Analysis
- Sparse and Compressive Sensing Techniques
- Functional Brain Connectivity Studies
- Neonatal and fetal brain pathology
- Image Retrieval and Classification Techniques
- Adversarial Robustness in Machine Learning
- Fetal and Pediatric Neurological Disorders
- Brain Tumor Detection and Classification
- Advanced MRI Techniques and Applications
- Advanced Image and Video Retrieval Techniques
- Advanced Image Processing Techniques
- Glioma Diagnosis and Treatment
- COVID-19 diagnosis using AI
- Image and Signal Denoising Methods
- Advanced Vision and Imaging
- Neural Networks and Applications
- Generative Adversarial Networks and Image Synthesis
- Video Coding and Compression Technologies
- Multimodal Machine Learning Applications
École de Technologie Supérieure
2016-2025
Université Paris-Saclay
2025
Centre National de la Recherche Scientifique
2023-2025
International Laboratory on Learning Systems
2024
Université de Montréal
2018-2024
Département d'Informatique
2022
Université du Québec à Montréal
2015-2020
Imagerie et Cerveau
2020
McGill University
2020
Ciena (Canada)
2020
Recently, dense connections have attracted substantial attention in computer vision because they facilitate gradient flow and implicit deep supervision during training. Particularly, DenseNet that connects each layer to every other a feed-forward fashion has shown impressive performances natural image classification tasks. We propose HyperDenseNet, 3-D fully convolutional neural network extends the definition of connectivity multi-modal segmentation problems. Each imaging modality path,...
Accurate segmentation of infant brain magnetic resonance (MR) images into white matter (WM), gray (GM), and cerebrospinal fluid is an indispensable foundation for early studying growth patterns morphological changes in neurodevelopmental disorders. Nevertheless, the isointense phase (approximately 6-9 months age), due to inherent myelination maturation process, WM GM exhibit similar levels intensity both T1-weighted T2-weighted MR images, making tissue very challenging. Although many efforts...
We address the problem of segmenting 3D multi-modal medical images in scenarios where very few labeled examples are available for training. Leveraging recent success adversarial learning semi-supervised segmentation, we propose a novel method based on Generative Adversarial Networks (GANs) to train segmentation model with both and unlabeled images. The proposed prevents over-fitting by discriminate between true fake patches obtained generator network. Our work extends current approaches,...
Emerging evidence suggests the presence of neuroanatomical abnormalities in subjects with autism spectrum disorder (ASD). Identifying anatomical correlates could thus prove useful for automated diagnosis ASD. Radiomic analyses based on MRI texture features have shown a great potential characterizing differences occurring from tissue heterogeneity, and identifying related to these differences. However, only limited number studies investigated link between image This paper proposes study grey...
Precise segmentation of bladder walls and tumor regions is an essential step towards non-invasive identification stage grade, which critical for treatment decision prognosis patients with cancer (BC). However, the automatic delineation in magnetic resonance images (MRI) a challenging task, due to important shape variations, strong intensity inhomogeneity urine very high variability across population, particularly on tumors appearance. To tackle these issues, we propose use deep fully...
Purpose Precise delineation of organs at risk is a crucial task in radiotherapy treatment planning for delivering high doses to the tumor while sparing healthy tissues. In recent years, automated segmentation methods have shown an increasingly performance various anatomical structures. However, this remains challenging like esophagus, which versatile shape and poor contrast neighboring For human experts, segmenting esophagus from CT images time‐consuming error‐prone process. To tackle these...
Alzheimer's disease (AD) is the most common form of dementia, causing progressive impairment memory and cognitive functions. Radiomic features obtained from brain MRI have shown a great potential as non-invasive biomarkers for this disease; however, their usefulness has not yet been explored individual regions. In paper, we hypothesize that distinct regions are affected differently by AD and, thus, shape or texture changes occurring in separate can be expressed different radiomic features....
This paper presents a novel set of image texture features generalizing standard grey-level co-occurrence matrices (GLCM) to multimodal data through joint intensity (JIMs). These are used predict the survival glioblastoma multiforme (GBM) patients from MRI data. The scans 73 GBM Cancer Imaging Archive in our study. Necrosis, active tumor, and edema/invasion subregions phenotypes segmented using coregistration contrast-enhanced T1-weighted (CE-T1) images its corresponding fluid-attenuated...
Background: Colorectal cancer (CRC) is the third most common among men and women. Its diagnosis in early stages, typically done through analysis of colon biopsy images, can greatly improve chances a successful treatment. This paper proposes to use convolution neural networks (CNNs) predict three tissue types related progression CRC: benign hyperplasia (BH), intraepithelial neoplasia (IN), carcinoma (Ca). Methods: Multispectral images thirty CRC patients were retrospectively analyzed. Images...
// Ahmad Chaddad 1, 2 , Christian Desrosiers Matthew Toews and Bassam Abdulkarim 1 Division of Radiation Oncology, McGill University, Montréal, Canada The Laboratory for Imagery, Vision Artificial Intelligence, Ecole de Technologie Supérieure, Correspondence to: Chaddad, email: ahmad.chaddad@mail.mcgill.ca Keywords: lung cancer; NSCLC; cancer staging; radiomics; texture features Received: May 30, 2017     Accepted: October 02, Published: November 01, 2017...
In recent years, approaches based on nonlocal self similarity and global structure regularization have led to significant improvements in image restoration. Nonlocal exploits the repetitiveness of small patches as a powerful prior reconstruction process. Likewise, is principle that objects represented by relatively portion pixels. Enforcing this structural information be sparse can thus reduce occurrence artifacts. So far, most restoration considered one these two strategies, but not both....
Predictors of patient outcome derived from gene methylation, mutation, or expression are severely limited in IDH1 wild-type glioblastoma (GBM). Radiomics offers an alternative insight into tumor characteristics which can provide complementary information for predictive models. The study aimed to evaluate whether models integrate radiomic, gene, and clinical (multi-omic) features together offer increased capacity predict outcome. A dataset comprising 200 GBM patients, Cancer Imaging Archive...