Robust affine point matching via quadratic assignment on Grassmannians
This addresses robust point cloud matching for applications like computer vision or robotics, but appears incremental as it builds on existing affine matching and QAP techniques.
The paper tackles the problem of affine registration of point clouds by introducing the RoAM algorithm, which minimizes the Frobenius distance on Grassmannians using an indefinite relaxation of the Quadratic Assignment Problem, and shows it is more robust to noise and point discrepancy than previous methods.
Robust Affine Matching with Grassmannians (RoAM) is a new algorithm to perform affine registration of point clouds. The algorithm is based on minimizing the Frobenius distance between two elements of the Grassmannian. For this purpose, an indefinite relaxation of the Quadratic Assignment Problem (QAP) is used, and several approaches to affine feature matching are studied and compared. Experiments demonstrate that RoAM is more robust to noise and point discrepancy than previous methods.