Recogniton Research Of Handwritten Music Symbol Based On Classroom Teaching Of Digital Music

With the rapid development of the computer technology, people pay more attention to the application of modern information technology in teaching, and a large number of educational software have gradually infiltrated into teaching of various subjects. However, because of the contradiction between the low-speed input of the traditional software and the high-speed processing of the computer has brought inconvenience to the teaching, there is an urgent need to turn traditional input of keyboard into the handwriting input in accordance with the people’s daily habits, which can enhance the classroom instruction efficiency.The main content of the paper is about the study of the recognition of on-line handwritten notes, it is based on the metaphor of "pen and paper" and the design idea of natural interaction with user centered, so as to provide a natural and efficient environment for teaching of music education. This paper mainly discusses and studies the following issues:(1)According to the geometric characteristics of the strokes of the notes designed, electing eight geometric features and one eight direction chain code which is suited for the strokes set. We select the appropriate threshold for geometric features by analyzing the data collection of the characteristics of the sample set. Meanwhile, it has also carried on the processing of duplicate removal for eight direction chain code of strokes, and selected more appropriate critical angle. These operations make full preparations for the subsequent recognition research.(2)It designs the recognizer in multi classifier fusion thought, and proposes single stroke classification method which is the combination of geometric features and eight direction chain code features. First, it rapidly conducts coarse classification for stroke in straight line rate feature; Then it conducts subdivisions for all kinds of inside by using the other geometric features and eight direction chain code, which combines minimum edit distance of the string matching theory and dynamic programming algorithm design techniques. Finally, through the validation of the experiment, the single stroke recognition rate reaches98.08%,which is higher than the Rubin algorithm that is used specially for single stroke recognition.(3)Unlike the handwritten Chinese characters, completing a handwritten note may be based on another note. According to the characteristic of the notes, it puts forward the method of strokes combine gradually. Through the semantic rules and positional relationship between the original input note and current stroke to determine whether they can be combined. The application of this method to a certain extent solve the problem of similarity and order between strokes.In a word, handwritten note of this thesis is more in line with people’s handwriting spectrum habits, providing users with a relatively free and efficient space for writing. For future extension and improvement of the algorithm has laid a good foundation.