A point pattern matching algorithm based on QR decomposition

A new method for point pattern matching is proposed in this paper. First, we show how the orthogonal-triangular decomposition (QR decomposition) can be used for point pattern matching. Second, in order to improve the matching accuracy, we propose a simple but robust point-matching method relying on finding the first K similar neighbors emerging from candidate matches, to embed the QR decomposition method within the framework of iterative matching. The proposed method, when applied to a wide experimental data has shown higher accuracy than the existing methods in this area.