Amir Nazemi

“The longer you look at an object, the more abstract it becomes, and, ironically, the more real.” - Lucian Freud

I am a Postdoctoral Researcher in the Systems Design Engineering department at the University of Waterloo, specializing in computer vision, machine learning, and artificial intelligence applications. My research focuses on developing innovative AI solutions for sports analytics, medical and nutrition data analysis and inverse problem.

Current Research

My current work involves sports analytics, particularly ice hockey broadcast video analysis, where I develop computer vision systems for:

• Homography estimation for rink-agnostic analysis
• Puck detection and tracking
• Player action recognition
• Game event detection and analysis

Research Background

I completed my Ph.D. in Systems Design Engineering at the University of Waterloo (2019–2023), specializing in Continual Learning-based Video Object Segmentation. My research focused on addressing the challenges of memory-efficient continual learning in long video sequences, particularly under conditions of domain shift. I developed novel methods that maintain segmentation performance over extended temporal spans while adhering to strict computational and memory constraints.

Industry Collaborations

I have actively contributed to several research collaborations between the University of Waterloo and industry partners:

• Stathletes Inc.
  Project: Ice hockey broadcast video analysis
  Role: University–industry liaison and primary contact point

• Moonrise Medical Inc.
  Project: Machine learning models for Doppler ultrasound data
  Role: Lead collaborator and main contact point between the university and the company

• Microsoft Office Media Group
  Project: Visual relation detection and semi-supervised video object segmentation
  Role: Ph.D. researcher contributing to joint academic–industry initiatives

Teaching & Mentorship

I am passionate about education and have taught graduate-level courses including Statistical Image Processing and Multidimensional Modeling. I currently supervise and mentor multiple graduate and undergraduate students, utilizing project management approaches like Kanban methods for academic research coordination.

Awards & Recognition

• Best Paper Award at the Linköping Hockey Analytics Conference (2025)
• CVPR CLVISION Challenge Finalist (2020)
• International Doctoral Student Award from University of Waterloo (2019-2023)
• Traffic Safety Award (2015)

My research has been published in top-tier venues including IEEE/CVF Conference on Computer Vision and Pattern Recognition, Neurips, IEEE Transactions on Intelligent Transportation Systems, and IEEE Access.

For more details about my research, publications, and experience, please explore the publications and CV sections of this website.