6D Pose Estimation from Single RGB Image

Estimating 6D object pose from a single RGB image using a backpropagatable PnP formulation.

Developed a method for estimating 6D object pose from a single RGB image using a differentiable (backpropagatable) Perspective-n-Point (PnP) formulation, enabling end-to-end training for pose estimation.

This project was part of my graduate coursework and research at the University of Michigan.

References