Computer-vision

FormNet: Visual Identification of Articulated Object Parts

FormNet: a deep neural network that identifies real-world articulation mechanisms (e.g., cabinets, drawers, etc) from a single RGB-D frame. (IROS 2021)

Camera-to-Robot Pose Estimation from a Single Image

Developed an approach for camera calibration from a single RGB image of a robot manipulator. This was my internship project at NVIDIA during Summer 2019.

Vision-Based Localization for Robotic Inspection of Nuclear Reactors

Using a pan-tilt-zoom camera, a remotely operated vehicle (ROV) was accurately localized within a nuclear reactor, providing greater utility and assurance for reactor inspection operations. This project was my M.S. in Robotics thesis research.