Max Robotics — Global Service Robot Directory + US Certification

About

Marwan Torki is a computer vision and machine learning researcher whose work spans human action recognition, motion analysis, and brain-computer interfaces. He is best known for his influential contributions to activity recognition from video and depth sensor data, particularly leveraging skeletal joint information to characterize human movement. His 2013 paper on the temporal hierarchy of covariance descriptors applied to 3D joint locations has accumulated over 530 citations, establishing a widely adopted framework for recognizing human actions in challenging real-world scenarios. Complementing this, his Histogram of Oriented Displacements (HOD) method for trajectory description of human joints garnered over 180 citations, further demonstrating his ability to design compact yet discriminative motion representations. Torki has also explored RGB-D sensing for object pose recognition, capitalizing on the growing availability of affordable depth cameras to advance robotic perception. More recently, his research has expanded into neurotechnology, investigating deep learning-driven electroencephalography-based brain-computer interfaces for robot control. Across these diverse threads, his work consistently addresses real-world applications in human-robot interaction, surveillance, and assistive technology, making him a notable contributor to embodied AI and perceptual computing.