Max Robotics — Global Service Robot Directory + US Certification

About

Brian Okorn is a roboticist whose research sits at the intersection of computer vision and manipulation, tackling the fundamental challenge of enabling robots to interact with objects in unstructured environments. His work centers on object rearrangement, pose estimation, and skill transfer, with a focus on generalization beyond trained scenarios. Okorn’s most impactful contribution is **IFOR: Iterative Flow Minimization for Robotic Object Rearrangement** (34 citations), an end-to-end method that uses iterative flow to accurately rearrange objects from vision alone—a critical capability for real-world robotics. He also developed **OSSID: Online Self-Supervised Instance Detection by (And For) Pose Estimation**, a novel approach that eliminates the need for retraining pose estimation models for each new object, and **TAX-Pose: Task-Specific Cross-Pose Estimation for Robot Manipulation**, which enables robots to transfer manipulation skills to unseen objects by focusing on task-specific pose relationships. Earlier in his career, Okorn applied robotics to security challenges, co-authoring work on autonomous tunnel exploration with unmanned ground vehicles. His contributions are shaping how robots perceive and manipulate the physical world, bridging the gap between demonstration and autonomous operation.