A5072033636- Radiomics and Machine Learning in Medical Imaging
- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- AI in cancer detection
- Hepatocellular Carcinoma Treatment and Prognosis
- Liver Disease Diagnosis and Treatment
- Domain Adaptation and Few-Shot Learning
- Medical Imaging Techniques and Applications
- Burn Injury Management and Outcomes
- MRI in cancer diagnosis
- Brain Metastases and Treatment
- Topic Modeling
- Pressure Ulcer Prevention and Management
- Brain Tumor Detection and Classification
- Metabolism, Diabetes, and Cancer
- Advanced Radiotherapy Techniques
- Medical Imaging and Analysis
- COVID-19 diagnosis using AI
- Veterinary Orthopedics and Neurology
- Generative Adversarial Networks and Image Synthesis
- Veterinary Equine Medical Research
- Advanced MRI Techniques and Applications
- Robotics and Sensor-Based Localization
- Advanced X-ray and CT Imaging
- Machine Learning and Algorithms
GDI Integrated Facility Services (Canada)
2022
Université de Montréal
2015-2022
Centre Hospitalier de l’Université de Montréal
2015-2017
École de Technologie Supérieure
2014-2017
Centre for Interdisciplinary Research in Rehabilitation
2017
Hôpital Saint-Luc
2017
Université du Québec à Montréal
2015
In this work, we report the set-up and results of Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with IEEE International Symposium on Biomedical Imaging (ISBI) 2017 Conferences Medical Image Computing Computer-Assisted Intervention (MICCAI) 2018. The image dataset is diverse contains primary secondary tumors varied sizes appearances various lesion-to-background levels (hyper-/hypo-dense), created collaboration seven hospitals research institutions. Seventy-five...
In this work, we report the set-up and results of Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with IEEE International Symposium on Biomedical Imaging (ISBI) 2017 Conferences Medical Image Computing Computer-Assisted Intervention (MICCAI) 2018. The image dataset is diverse contains primary secondary tumors varied sizes appearances various lesion-to-background levels (hyper-/hypo-dense), created collaboration seven hospitals research institutions. Seventy-five...
Deep learning is a class of machine methods that has been successful in computer vision. Unlike traditional require hand-engineered feature extraction from input images, deep learn the image features by which to classify data. Convolutional neural networks (CNNs), core for imaging, are multilayered artificial with weighted connections between neurons iteratively adjusted through repeated exposure training These have numerous applications radiology, particularly classification, object...
OBJECTIVE This study determined the effects of insulin versus liraglutide therapy on liver fat in patients with type 2 diabetes inadequately controlled oral agents therapy, including metformin. RESEARCH DESIGN AND METHODS Thirty-five metformin monotherapy or combination other antidiabetic medications were randomized to receive glargine for 12 weeks. The proton density fraction (PDFF) was measured by MRS. mean PDFF, total volume, and index MRI. Student t test, Fisher exact repeated-measures...
The purpose of this paper is to describe a semiautomated segmentation method for the liver and evaluate its performance on CT-scan MR images.First, an approximate 3-D model initialized from few user-generated contours globally outline shape. then automatically deformed by Laplacian mesh optimization scheme until it precisely delineates patient's liver. A correction tool was implemented allow user improve satisfaction.The proposed tested against 30 CT-scans SLIVER07 challenge repository 20...
Purpose To assess the agreement between published magnetic resonance imaging (MRI)‐based regions of interest (ROI) sampling methods using liver mean proton density fat fraction (PDFF) as reference standard. Materials and Methods This retrospective, internal review board‐approved study was conducted in 35 patients with type 2 diabetes. Liver PDFF measured by spectroscopy (MRS) a stimulated‐echo acquisition mode sequence MRI multiecho spoiled gradient‐recalled echo at 3.0T. ROI reported...
Background Detection of brain metastases (BM) and segmentation for treatment planning could be optimized with machine learning methods. Convolutional neural networks (CNNs) are promising, but their trade‐offs between sensitivity precision frequently lead to missing small lesions. Hypothesis Combining volume aware (VA) loss function sampling strategy improve BM detection sensitivity. Study Type Retrospective. Population A total 530 radiation oncology patients (55% women) were split into a...
Liver volumetry is considered to be an accurate indicator of hepatic function and a prognostic in surgery planning. Despite many years research, automated liver segmentation remains open challenge manual still widely used clinically although it time-consuming tedious. In this paper we propose novel semi-automated method based on deformable models independent training data. First, initial shape the generated by variational interpolation from few user-generated contours. A template-matching...
Current deep learning based text classification methods are limited by their ability to achieve fast and generalization when the data is scarce. We address this problem integrating a meta-learning procedure that uses knowledge learned across many tasks as an inductive bias towards better natural language understanding. Based on Model-Agnostic Meta-Learning framework (MAML), we introduce Attentive Task-Agnostic (ATAML) algorithm for classification. The essential difference between MAML ATAML...
This paper reports a novel approach to 3D kidney segmentation from single prior shape in magnetic resonance imaging (MRI) datasets. The proposed method is based on hierarchic surface deformation algorithm, generate pre-personalized model, followed by an anamorphing extract the capsule. Accuracy and precision are assessed comparing our over 20 reconstructions segmented manually 3 different observers native MRI images. experimental results show volumetric overlap error of 6.39±2.47%, relative...
In clinical practice, knee MRI sequences with 3.5~5 mm slice distance in sagittal, coronal, and axial planes are often requested for the examination since its acquisition is faster than high-resolution sequence a single plane, thereby reducing probability of motion artifact. order to take advantage three from different planes, 3D segmentation method based on combination models obtained proposed this paper. method, sub-segmentation respectively performed image coordinate system. With each...
In this paper, we study the influence of both long and short skip connections on Fully Convolutional Networks (FCN) for biomedical image segmentation. standard FCNs, only are used to features from contracting path expanding in order recover spatial information lost during downsampling. We extend FCNs by adding connections, that similar ones introduced residual networks, build very deep (of hundreds layers). A review gradient flow confirms a FCN it is beneficial have connections. Finally,...
Abstract The burden of detection and segmentation brain metastases (BM) for treatment planning response assessment has been found to be alleviated by machine learning methods. However, tracking individual lesions over time remains tedious would benefit from automated assistance complex cases. We developed a software solution combining an AI-based BM method with pairing algorithm allowing track across longitudinal scans. proposed comprises two steps: identifying in each scan identified...
In this paper, we introduce a simple, yet powerful pipeline for medical image segmentation that combines Fully Convolutional Networks (FCNs) with Residual (FC-ResNets). We propose and examine design takes particular advantage of recent advances in the understanding both Neural as well ResNets. Our approach focuses upon importance trainable pre-processing when using FC-ResNets show low-capacity FCN model can serve pre-processor to normalize input data. our pipeline, use FCNs obtain normalized...