In China note processing is a necessary work for finance and much workload is emphasized
on it. The feature recognition of notes is an important work of financial management information and
office automation. Feature extraction can quickly locate the notes' size and shape. It can quickly locate
the position of the critical information, and improve the speed of computer recognition. In this way the
financial management efficiency can be improved greatly. By SURF algorithm the image feature
points are detected and Flann method is used to match feature points. In this way we extract the region with corresponding
characteristics of template collection after clustering. There are also many patents on these intelligence methods. Experiments
show that our algorithm has better matching rate, and it has certain robustness to the rotated images.
Keywords: Notes, feature matching, clustering, geometric characteristics.
Rights & PermissionsPrintExport