Identification method for handwritten form Tibetan characters

Abstract

The invention puts forward an identification method for handwritten form Tibetan characters, and can fully utilize the online and offline information of the handwritten form Tibetan characters and improve the identification rate of the handwritten form Tibetan characters. The method comprises the following implementation steps that: (1) establishing a Tibetan character dictionary; (2) establishinga handwritten form Tibetan character sample library; (3) obtaining the online characteristic of each sample in the handwritten form Tibetan character sample library; (4) obtaining the offline characteristic of each sample in the handwritten form Tibetan character sample library; (5) obtaining the characteristic vector of the Tibetan character sample which does contain noise; (6) obtaining a training sample set and samples to be identified; (7) obtaining the characteristic template of the Tibetan character sample; and (8) classifying the Tibetan character samples in the samples to be identified. The identification method can be used for identifying the handwritten form Tibetan characters input by a mobile terminal.

Description

technical field [0001] The invention belongs to the technical field of character pattern recognition, and relates to a handwritten Tibetan character recognition method, which can be used to recognize the handwritten Tibetan character input by a mobile terminal. Background technique [0002] The Tibetan language has a long history, and the domestic population is about 8 million. At present, the English and Chinese recognition technology has matured and is widely used in various fields, while the Tibetan language recognition technology is still in its infancy, and the results are relatively few. With the popularity of various mobile devices, handwriting input has become an important way of human-computer interaction. Tibetan handwriting recognition not only has important social significance, but also has broad market prospects. [0003] The Tibetan language structure can be divided into two parts: consonants and vowels. These letters appear in different positions of the Tibeta...

AI Technical Summary

Problems solved by technology

(2) The number of Tibetan syllables is too large, which will not only cause great trouble for recognition, but also extremely difficult for data collection based on existing conditions

At present, in the existing technology, the recognition of Tibetan is completed by using a single offline or online recognition technology, which cannot make full use of the offline and online information of handwritten Tibetan, resulting in a low recognition rate of handwritten Tibetan. The handwritten Tibetan recognition method using two kinds of information has no relevant literature or patent reports

Images