Max Robotics — Global Service Robot Directory + US Certification

About

Kuan-Hui Lee is a leading researcher in computer vision and robotics, specializing in spatiotemporal reasoning, trajectory forecasting, and self-supervised 3D perception. His most influential work, "Spatiotemporal Relationship Reasoning for Pedestrian Intent Prediction" (185 citations), introduces a novel framework that models the dynamic interactions between pedestrians and their environment to anticipate future actions—a critical capability for safe autonomous navigation. Lee has also advanced self-supervised learning for 3D perception with "Learning Optical Flow, Depth, and Scene Flow Without Real-World Labels" (54 citations), enabling robots to learn depth and motion estimation directly from raw video without costly annotations. In "Heterogeneous-Agent Trajectory Forecasting Incorporating Class Uncertainty" (38 citations), he addresses the challenge of predicting the future behavior of diverse agents (e.g., pedestrians, cyclists, vehicles) by accounting for uncertainty in both their states and semantic classes. Additionally, Lee's work on counterfactual analysis in simulation ("Discovering Avoidable Planner Failures of Autonomous Vehicles," 6 citations) provides a rigorous method for identifying and mitigating safety-critical failures in autonomous vehicle planners. His research has been published at top venues including CVPR and IROS, and his contributions are shaping the next generation of perception and prediction systems for autonomous robots.