Method and system for recognizing a handwritten character

A system for recognizing a handwritten character, comprising:

• a storage module, being configured to store character classes, a coarse classification template and a fine classification template corresponding to the character classes, and an eigen transformation matrix calculated according to the Fisher criteria;
• a first signal gathering module, being configured to receive a handwritten character input signal from a user and gather a discrete coordinate sequence of trajectory points of the handwritten character;
• a first normalizing module, being configured to transform the discrete coordinate sequence of the trajectory points of the handwritten character that is gathered by the first signal gathering module into another discrete coordinate sequence so as to adjust the size, the shape and coordinate values of a center of gravity of the handwritten character to obtain a normalized coordinate sequence of the character;
• a first eigenvalue extracting module, being configured to, according to the normalized coordinate sequence, decompose vector line sections formed by all adjacent trajectory points of the handwritten character into eight standard directions to calculate a multi-dimensional eigenvector of the handwritten character;
• a coarse classification module, being configured to select a part of eigenvalues from the multi-dimensional eigenvector of the handwriting character inputted by the user, match the handwritten input character with the coarse classification template in the storage module respectively to select a plurality of the most similar sample centers from the coarse classification template, and use character classes corresponding to the plurality of sample centers as candidate character classes;
• a fine classification module, being configured to perform eigen transformation on the multi-dimensional eigenvector of the inputted character, select sample centers of the candidate character classes from the fine classification template for matching with the eigen-transformed inputted character respectively, and determine the most similar character classes among the candidate character classes for the user to select;
• and
• a display module, being configured to display the most similar character classes outputted by the fine classification module for the user to select.