Research on the handwriting identification based on the energy resolution

Feature extraction and identification is the key part of handwriting identification. In this paper, the testing handwriting and the training handwriting samples were preprocessed, and the characters were divide into the single ones, and then the single handwriting image was divided into M*N continuous meshes to extract the average gray level, which marked energy distribution. The eigenvector was structured, the comparability measurement function was constructed to match the character, the number of mesh is changed dynamically, and the optimal match result is searched. The experiments proved that the method improves the precision of handwriting recognition effectively.