Greater Pittsburgh Region
Computer vision research Specialties: Computer Vision, Machine Learning and Accessibility
Computer Vision, Machine Learning, Human-Robot Interaction. Physics-based egocentric pose estimation, activity & pose forecasting, vision-based collision avoidance, 3D object detection, 3D multi-object tracking, neural network compression, multi-agent imitation learning and modeling, camera calibration, co-manipulation, gaze prediction, deep generative models, multi-agent forecasting, synthetic data for domain adaptation, self-supervised learning, action recognition, anomaly detection, multi-modal localization, state estimation, non-line of sight imaging for human activity understanding, image and video accessibility on social media, automatic photo cropping, automatic photo acquisition with UAVs.
Wearable cameras, First-person vision, Activity forecasting and prediction, Inverse optimal control, Pedestrian detection by synthesis, Vision for navigation for Blind people, Vision for upper-extremity stroke rehabilitation ... Deep Learning.
Activity Forecasting
Gazed-based activity analysis
SAM 3D Body Lead (https://github.com/facebookresearch/sam-3d-body) Ego-Exo4D Co-Lead (https://ego-exo4d-data.org) Ego4D Co-Lead (https://ego4d-data.org)
Director of the Masters of Science in Computer Vision program