Feature extraction of Liquid Drop Fingerprint based on Bezier curve fitting

In order to effectively characterize the Liquid Drop Fingerprint (LDF) of different liquids, a new method based on Bezier curve fitting is put forward. The fitting curve is calculated by the least square method. To decrease the fitting error and the feature dimension, the Bezier curve order should be reasonably chosen, combining the particularity of LDF. Theoretical analysis and experimental results show that the eigenvector can reflect the typical differences among different kinds of liquids, and the recognition rate within current test samples is reached up to 100%.