Machine Learning and Computer Vision

Research Description 

We are developing machine learning algorithms to automatically track, model, and control multibody dynamic systems. Our efficient algorithms are deployed on mobile devices and control hardware units. Applications include markerless tracking of human movements, automatic extraction of anatomical parameters and landmarks from medical images, control of autonomous vehicles, automotive powertrain models, human muscle torque estimation, and environment recognition systems for lower-limb exoskeletons and prostheses.

Student Researchers 

Ali Nasr
AliAsghar MohammadiNasrabadi
Kevin Zhu
Nishad Rajmalwar
Hisham Mohammad
Chris Shum 
Yuan Lin
William McNally (Alumnus)
Arash Hashemi (Alumnus)
Brokoslaw Laschowski (Alumnus)

Keywords and Themes 

• Deep Learning and Computer Vision
• Medical Imaging
• Reinforcement Learning
• Autonomous Vehicles
• Environment Recognition Systems
• Machine Learning-driven Exoskeleton Control


Related Publications 

• MohammadiNasrabadi A, McNally W, Moammer G, McPhee J (2022). Automatic extraction of spinopelvic parameters using deep learning to detect landmarks as objects. In Medical Imaging with Deep Learning (MIDL), 2022.
• Nasr A and McPhee J. (2022). Biarticular MuscleNET: A machine learning model of biarticular muscles. In Proceedings of the North American Congress of Biomechanics (NACOB).
• Nasr A, Inkol K A, Bell S, and McPhee J. (2021). InverseMuscleNET: Alternative machine learning solution to static optimization and inverse muscle model. Frontiers in Computational Neuroscience. DOI: 10.3389/fncom.2021.759489
• Nasr A, Bell S, He J, Whittaker R L, Jiang N, Dickerson C R, and McPhee J. (2021). MuscleNET: Mapping electromyography to kinematic and dynamic biomechanical variables. Journal of Neural Engineering. vol. 18, no. 4, p.0460d3. DOI: 10.1101/2021.07.07.451532.
• Lin Y, McPhee J, and Azad NL. (2019). Comparison of Deep Reinforcement Learning and Model Predictive Control for Adaptive Cruise Control. arXiv:1910.12047  
• McNally W, Wong A, and McPhee J. (2019). STAR-Net: Action Recognition using Spatio-Temporal Activation Reprojection. arXiv:1902.10024.
• McNally W, Vats K, Pinto T, Dulhanty C, McPhee J, and Wong A. (2019). GolfDB: A Video Database for Golf Swing Sequencing. IEEE Conference on Computer Vision and Pattern Recognition (CVPR). arXiv:1903.06528.