I have a passion for designing machine learning and artificially intelligent models for medicine. We can do more to improve patient care and treatment by leveraging A.I. and I’m excited to be part of that change.

Here you’ll find some of the projects that I have tackled over my career as a graduate student, postdoctoral fellow and now industry researcher. My publications range from tumor segmentation algorithms to neural networks which learns from weak labels. Take a look around!

Shazia Akbar, PhD



Altis Labs, Toronto (Canada)2019 – 2021
Machine Learning Engineer
Sunnybrook Research Institute (Canada)
Medical Biophysics, University of Toronto
Vector Institute
2016 – 2019
Postdoctoral Fellow
New York University, School of Medicine (USA)2015 – 2016
Postdoctoral Fellow
Toshiba Medical Visualization Systems (UK)2014 – 2014
Image Analysis Research Intern
NCR Corp. (UK)
Technology and Research Intern2010 – 2011


University of Dundee (UK)2011 – 2015
PhD: Tumor Localisation in Histopathology Images
University of Dundee (UK)2008 – 2011
BSc Applied Computing (Hons)

Download Resume



Here is a snapshot of some of the methods I have developed:

Multiple Instance LearningTransition Module
Technique for training neural networks with weak or coarse labels grouped together in “bags”A method for preventing overfitting in 2D convolutional neural networks by gradually reducing dimensionality

Tumor Cellularity Assessment

Superpixel Classification
Fully automated neural network that assesses cellularity in digitized histology images for tumor burden assessmentAlgorithm for classifying superpixels using features encapsulating contextual information


My most recent publications are shown here but a full list is also available.

Torres, Felipe S; Akbar, Shazia; Raman, Srinivas; Yasufuku, Kazuhiro; Schmidt, Carola; Hosny, Ahmed; Baldauf-Lenschen, Felix; Leighl, Natasha B

End-to-End Non-Small-Cell Lung Cancer Prognostication Using Deep Learning Applied to Pretreatment Computed Tomography Journal Article

JCO Clin Cancer Inform , 5 , pp. 1141-1150, 2021.

Abstract | Links | BibTeX

Petrick, Nicholas; Akbar, Shazia; Cha, Kenny H; Nofech-Mozes, Sharon; Gavrielides, Berkman Sahiner Marios A; Kalpathy-Cramer, Jayashree; Drukker, Karen; Martel, Anne L; Group, BreastPathQ Challenge

SPIE-AAPM-NCI BreastPathQ challenge: an image analysis challenge for quantitative tumor cellularity assessment in breast cancer histology images following neoadjuvant treatment Journal Article

SPIE Medical Imaging, 8 (3), 2021.

Abstract | Links | BibTeX

Akbar, Shazia; Peikari, Mohammad; Salama, Sherine; Panah, Azadeh Yazdan; Nofech-Mozes, Sharon; Martel, Anne L

Automated and Manual Quantification of Tumour Cellularity in Digital Slides for Tumour Burden Assessment Journal Article

Scientific Reports, 2019.

Abstract | Links | BibTeX

Akbar, Shazia; Peikari, Mohammad; Salama, Sherine; Nofech-Mozes, Sharon; Martel, Anne L

The transition module: A method for preventing overfitting in convolutional neural networks Journal Article

Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 7 , 2019.

Abstract | Links | BibTeX

Akbar, Shazia; Martel, Anne L

Cluster-based learning from weakly labeled bags in digital pathology Conference

Machine Learning for Health Workshop, NeurIPS 2018, 2018.

Abstract | Links | BibTeX

Full list