Identification method for handwriting latin letter

Abstract

The invention discloses a recognition method for free handwritten Latin letters, which has the advantages that the method comprises the following steps: considering multiple writing styles of Latin letters, extracting various stable characteristics of the Latin letters at different writing styles, and classifying a Latin letter set once by using one characteristic, so as to classify the whole standard Latin letter set into a plurality of mutually intersected small Latin letter sets as rough classified candidate letter sets during recognition; and when input handwritten Latin letters are recognized, extracting classification characteristics of the input handwritten Latin letters to match the characteristics layer by layer till the rough classified candidate letter sets matched with the letters are obtained. The method ensures recognition rate through extraction of the various stable characteristics, finds out a regular search path layer by layer from layer-by-layer classification of the letter sets and layer-by-layer characteristic match during recognition to final rough classified candidate letter sets when the input handwritten Latin letters are recognized, and improves recognition speed.

Description

technical field [0001] The invention relates to a handwriting recognition method, in particular to a free handwritten Latin letter recognition method. Background technique [0002] With the rapid development of information technology, various handwritten character recognition technologies continue to deepen, and various commercial recognition systems are put into the market. At present, research hotspots are focused on the development of various free handwritten letter recognition systems, among which free handwritten Latin letter recognition system is also one of the research hotspots. Although the Latin letter recognition character set is very small, some Latin letters have very little structure and shape difference, and are easily confused with each other. For example, when freely handwriting Latin letters such as a, b, q, p, n, and h, the differences between them It is small, which has a great impact on the recognition rate of the free handwritten Latin alphabet recogni...

AI Technical Summary

Problems solved by technology

The recognition process of this recognition system, whether it is a first-level feature or a second-level feature, depends on the extraction and division of strokes, but many free handwritten Latin letters such as a, b, q, p, n, h, etc., have relatively vague stroke features , and there are more arc parts, so the stroke feature is selected to recognize Latin letters, which greatly reduces the recognition rate

Images