Using Feature Spatial Order in Progressive Image Feature Matching

Image feature matching is a very important and fundamental task in computer vision. In this paper, a spatial-order based progressive feature matching framework is proposed. With the model of spatial order, the searching space is partitioned into many intervals with each interval associated with a probability that a correct match is occurred in this interval. Using this information, many incorrect features could be filtered out and only the survived features are passed for subsequent matching. As the features are progressively matched, the model of spatial order is also progressively updated and the lengths of partitioned intervals are further shortened to filter out more features. To demonstrate the feasibility of proposed system, a series of experiments were conducted. A standard benchmark image data set was used to test the proposed system and the results showed that the proposed framework can indeed produce more efficient and accurate feature matching compared with traditional brute force technique.