MIMLab at St. Michael's Hospital

where applied machine intelligence meets medical data

to enhance patient care

About Us

MIMLab is focused on medical imaging processing, big medical data analytics, and theoretical machine learning. Our expertise bridges the gap between data warehouses, modern algorithms, high performance computing, and better outcomes for patients, better and more cost-effective workflows, and product development. We have an active relationship with tech transfer at St. Michael's Hospital, MaRS and a variety of commercial interests.

News

  • April 2018 - NVIDIA granted a GPU to Dr. Djeven Deva for research in machine learning. Congratulations!

  • April 2018 - MIMLab's interview with Canadian Healthcare Technology Magazine: Link to the article-page 10

  • March 2018 - NVIDIA granted a GPU to Dr. Aditya Bharatha for research in machine learning. Congratulations!

  • February 2018 - The MIMLab welcomes Jiankun Wang aboard!

  • January 2018 - Our paper "IMAGE AUGMENTATION USING RADIAL TRANSFORM FOR TRAINING DEEP NEURAL NETWORKS" is accepted at the 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (IEEE ICASSP). ICASSP 2018 has received 2830 paper submissions, the second largest number of paper submissions for an ICASSP.

  • January 2018 - Our paper "GENERALIZATION OF DEEP NEURAL NETWORKS FOR CHEST PATHOLOGY CLASSIFICATION IN X-RAYS USING GENERATIVE ADVERSARIAL NETWORKS" is accepted at the 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (IEEE ICASSP). ICASSP 2018 has received 2830 paper submissions, the second largest number of paper submissions for an ICASSP.

  • August 2017 - Our paper "A Convolutional Neural Network for Search Term Detection" is accepted for presentation at 2017 IEEE 28th Annual International Symposium on Personal, Indoor, and Mobile Radio Communications (IEEE PIMRC).

  • June 2017 - The MIMLab is delighted to announce the pending installation of a GE Discovery CZT camera for nuclear cardiology. This camera will enable a set of projects on multiple body systems with an emphasis on big data and high performance GPU computing. Some of these projects include new quantitative approaches to myocardial perfusion imaging and myocardial function.

  • May 2017 - The MIMLab is delighted to announce the installation of a Digirad ERGO cesium iodide scintillation camera in the Deparmtent of Functioanl and Molecular Imaging. This new resource enables new applications including first pass myocardial imaging and molecular breast imaging.

  • May 2017 - The paper ``Training and Validating a Deep Convolutional Neural Network for Computer-Aided Detection and Classification of Abnormalities on Frontal Chest Radiographs" is accepted for publication at Investigative Radiology journal.

  • March 2017 - The MIMLab is delighted to announce the promotion of Dr. Joe Barfett to assistant professor at the University of Toronto. Congratulations Joe!

  • April 2016 - The MIMLab is deighted to announce the recruitment of Hojjat Salehinejad to the lab. We wish you tremendous succcess with your graduate work!

  • March 2016 - The MIMLab is delighted to announce the installation of a development workstation including two NVIDIA GPU Titans, an iMAC, and GPU enabled PC to the lab's core resources. Interested parties are invited to contact the team for processing time.

  • February 2016 - NVIDIA has donated two GPU Titan cards to MIM Lab. Thanks NVIDIA!

  • February 2016 - The MIM lab is delighted to introduce the GE Revolution voumetric CT system to the Department of Medical Imaging's core resources. This exciting new camera system enables a variety of new vascular applications with an emphasis on big data. Each acquisiton creates in excess of 8000 images, enabling characaterization of physiologic processes at exceptional spatial and temporal resolution.

Our Team

We are a network of physicians, scientists, engineers, clinician scientists and clinician engineers that aim to use modern image processing, statistical learning, machine learning, deep learning and high performance computing tools to solve medical problems.


Dr. Tim Dowdell

MD, CCFP, FRCPC

Associate Professor
Radiologist-in-Chief

St. Michael's Hospital

University of Toronto

Dr. Joseph Barfett

FRCPC, MD, MSc, BESc

Assistant Professor
Nuclear Medicine

St. Michael's Hospital

University of Toronto

Dr. Errol Colak

HMD, FRCPC, HBSc

Assistant Professor
Abdominal Imaging

St. Michael's Hospital

University of Toronto

Dr. Aditya Bharatha

BSc, MD, RCPC, FRCPC, DABR

Assistant Professor
Neuroradiologist

St. Michael's Hospital

University of Toronto


Prof. Shahrokh Valaee

PhD, P.Eng, FEIC

Professor

Department of Electrical & Computer Engineering

University of Toronto

Dr. Bruce Gray

MD, FRCPC, BSc

Assistant Professor
Neuroradiologist

St. Michael's Hospital

University of Toronto

Dr. Djeven Deva

FRCR, MRCSI, MB, BCh, BAO

Assistant Professor
Cardiothoracic Imaging

St. Michael's Hospital

University of Toronto

Hojjat Salehinejad

Lead Data Scientist
St. Michael's Hospital

PhD Candidate
Department of Electrical & Computer Engineering
University of Toronto


Jiankun Wang

MASc Candidate

Department of Electrical & Computer Engineering

St. Michael's Hospital

University of Toronto

Sumeya Naqvi

HBSc




St. Michael's Hospital

University of Toronto

Aren Mnatzakanian

HBSc Candidate




St. Michael's Hospital

University of Toronto

Hui-Ming Lin

HBSc Candidate




St. Michael's Hospital

University of Toronto


Rebecca Benjamin

HBSc Candidate




St. Michael's Hospital

McMaster University

Natalija Vagners

HBSc Candidate




St. Michael's Hospital

Ryerson University

Maya Czerwoniak

HBSc Candidate




St. Michael's Hospital

McMaster University

Research


Deep Learning

Medical Image Processing

Natural Language Processing in Medicine

Machine Learning in Neuroscience

Some of the recent projects:

  • Breast Density Estimation and Cancer Detection

    Breast density is one of the main risk factors for breast cancer. In this project, we have developed a novel approach to breast density estiamtion and cancer detection using deep neural networks (Current model is our previous proposed solution). More updates to come...

    Loading
    Links: Paper Code Data
  • Publications

    • Salehinejad H, Valaee S, Dowdell T, Colak E, Barfett J. Generalization of Deep Neural Networks for Chest Pathology Classification in X-Rays Using Generative Adversarial Networks. 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (IEEE ICASSP). Link

    • Salehinejad H, Valaee S, Dowdell T, Barfett J. Image Augmentation using Radial Transform for Training Deep Neural Networks. 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (IEEE ICASSP).Link

    • Alcaide-Leon P., Dufort P., Geraldo A.F., Alshafai L., Maralani P.J., Spears J. and Bharatha A. Differentiation of Enhancing Glioma and Primary Central Nervous System Lymphoma by Texture-Based Machine Learning. American Journal of Neuroradiology, 38(6), pp.1145-1150, 2017.Link

    • Salehinejad H, Baarbe J, Sankar S, Barfett J, Colak E, Valaee S. Recent Advances in Recurrent Neural Networks. arXiv preprint arXiv:1801.01078, 2017.Link

    • Salehinejad H, Barfett J, Valaee S, Colak E, Mnatzakanian A, Dowdell T. Interpretation of Mammogram and Chest Radiograph Reports using Deep Neural Networks. arXiv preprint arXiv:1708.09254, 2017.Link

    • Salehinejad H, Barfett J, Arabi P, Valaee S, Colak E, Gray B, Dowdell T. A Convolutional Neural Network for Search Term Detection. Personal, Indoor, and Mobile Radio Communication (PIMRC), 2017 IEEE 25th Annual International Symposium on. IEEE, 2017.Link

    • Cicero M, Bilbily A, Colak E, Dowdell T, Gray B, Perampaladas K, Barfett J. Training and validating a deep convolutional neural network for computer-aided detection and classification of abnormalities on frontal chest radiographs. Investigative radiology. 2017 May 1;52(5):281-7.

    Data

    Below is a collection of fine datasets related to medical imaging available on the Internet for use of researchers. We are working hard with the privacy office of St. Michael's hospital to be able to release our anonymized datasets for reseach community use.


    • Bone Age images used in the 2017 RSNA Bone Age Challenge competition - Stanford University Link
    • Labeled Musculoskeletal Radiographs Normal/Abnormal (50k images) - Stanford University Link
    • CBIS-DDSM (Curated Breast Imaging Subset of DDSM) - Cancer Imaging Archive Link
    • DDSM: Digital Database for Screening Mammography - University of South Florida Link

    Softwares

    Here is a collection of softwares and codes.

    Partners

    We are grateful to our collaborators:






    Join Us

    We always welcome brilliant, passionate, and dedicated individuals to MIMLab. Just send an email to salehinejadh-at-smh.ca!

    Get in touch


    Main Office
    St. Michael's Hospital
    Medical Imaging Department
    7th Floor - 193 Young St.
    Toronto, ON
    Canada
    Tel: (416) 360-4000
    Email: salehinejadh at smh dot ca