Non-rigid Image Feature Matching by Structure Constraints

In this paper, we propose a non-rigid feature matching approach for image registration. The non-rigid feature matching approach is formulated as a maximum likelihood (ML) estimation problem. The feature points of one image are represented by Gaussian mixture model (GMM) centroids, and are fitted to the feature points of the other image by moving coherently to encode the global structure. We constructed two local structure descriptors of connectivity matrix and Laplacian coordinate to preserve the local structure of these feature points. Furthermore, the expectation maximization (EM) algorithm is applied to solve for this ML problem. Experiments on public datasets and real images demonstrate that the proposed approach has better performance than current state-of-the-art methods.